post_disc_result_prepare.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils.ZR_utils import (get_matches_df, post_disc, plot_match_stats, change_post_disc_thr,\n",
    "                            get_nodes_df, get_clusters_list, get_matches_all)\n",
    "from utils.eval import evaluate\n",
    "from utils.helper_fncs import save_obj, load_obj\n",
    "\n",
    "from utils.db_utils import gold_fragments_df_for_signer\n",
    "from utils.feature_utils import (op_100_meancenter)\n",
    "\n",
    "import glob\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "label_dict = {'coverage': 'Coverage',\n",
    "                'ned':'NED', \n",
    "              'purity': 'Purity',\n",
    "              'avg_purity': 'Average Purity', \n",
    "              'clus_purity':'Cluster Purity', \n",
    "              'grouping_P': 'Grouping Precision',\n",
    "              'grouping_R': 'Grouping Recall',\n",
    "                'grouping_F': 'Grouping F-score',\n",
    "              'clus_purity_inv': 'Inverse Cluster Purity'}\n",
    "\n",
    "def slct_res(res,sel):\n",
    "    tmp = []\n",
    "    for i, r in enumerate(res):\n",
    "        if i not in sel: tmp.append(r)\n",
    "    return tmp\n",
    "\n",
    "\n",
    "def plot_curve(ax, results, ax_dict, marker, label, annot_freq=2):\n",
    "    # results: list of score dicts\n",
    "    data = np.zeros((len(results), 2))\n",
    "    annots = []\n",
    "    for i,score in enumerate(results):\n",
    "        data[i,0] = score[ax_dict['x']]\n",
    "        data[i,1] = score[ax_dict['y']]\n",
    "        annots.append(score['dtw'])\n",
    "\n",
    "    ax.plot(data[:,0], data[:,1], marker, label=label)\n",
    "    ax.set_xlabel(label_dict[ax_dict['x']])\n",
    "    ax.set_ylabel(label_dict[ax_dict['y']])\n",
    "    for i,ann in enumerate(annots):\n",
    "        \n",
    "        if (i%annot_freq != 0) or (annot_freq==-1): continue\n",
    "        ax.annotate(str(ann), \n",
    "             xy=(results[i][ax_dict['x']], results[i][ax_dict['y']]),  \n",
    "             xycoords='data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Signer03']\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'exp_name' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-10d4d04f2875>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     33\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mname_root\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexp_name\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     34\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 35\u001b[0;31m \u001b[0;32mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mexp_name\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'exp_name' is not defined"
     ]
    }
   ],
   "source": [
    "exp_root = '/home/korhan/Desktop/zerospeech2017/track2/src/ZRTools/exp/'\n",
    "\n",
    "exp_ = 'corAE_op100_s4s5_64_iter0_self_tanh_Signer07'\n",
    "#exp_ = 'Signer01_Signer08_c3_right_PCA50_10_3_02_7_05_055_11_06_03'\n",
    "#exp_ = 'corAE_op100_s4s5_64_iter0_self_tanh_Signer01'\n",
    "#exp_ = 'corAE_op100_s4s5_64_iter0_self_tanh_Signer04_Signer05_Signer0950_10_35_02_7_05_06_9_06_032'\n",
    "#exp_ = 'corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_4_02_7_05_06_7_06_032'\n",
    "#exp_ = 'Signer07op10050_10_5_02_7_05_065_13_06_04'\n",
    "#exp_ = 'corAE_op100_s4s5_64_iter0_self_tanhSigner0350_10_4_02_7_05_06_9_06_03'\n",
    "#exp_ = 'Signer03_Signer07_c3_right_PCA50_10_3_02_7_05_055_11_06_03'\n",
    "exp_ = 'Signer03_c3_right_PCA50_10_3_02_7_05_06_13_06_04'\n",
    "#exp_ = 'op100_Signer03_Signer0750_10_4_02_7_05_06_9_06_03'\n",
    "#exp_ = 'corAE_op100_s4s5_64_iter0_self_tanh_Signer03_Signer0750_10_4_02_7_05_06_9_06_032'\n",
    "#exp_ = 'op100_Signer08'\n",
    "#exp_ = 'corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032'\n",
    "\n",
    "signer_id = ['Signer0'+ str(i) for i in [8]]\n",
    "#signer_id = ['Signer0'+ str(i) for i in [3,7]]\n",
    "signer_id = ['Signer0'+ str(i) for i in [3]]\n",
    "\n",
    "print(signer_id)\n",
    "\n",
    "def full_exp_path(exp_):\n",
    "    exp_root = '/home/korhan/Desktop/zerospeech2017/track2/src/ZRTools/exp/'\n",
    "\n",
    "    try:\n",
    "        name_root = glob.glob(exp_root + 'zrroot0_{}*/exp/*'.format(exp_))[0]\n",
    "        exp_name = '/'.join(name_root.split('/')[-3:])[:-6]\n",
    "    except:\n",
    "        name_root = glob.glob(exp_root + '{}*/'.format(exp_))[0]\n",
    "        exp_name = glob.glob(exp_root + '{}*/'.format(exp_))[0].split('/')[-2][:-6]\n",
    "    \n",
    "    return name_root, exp_name\n",
    "    \n",
    "print(exp_name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Read 26 matches\n",
      "Read 17 matches\n",
      "Read 14 matches\n",
      "Read 23 matches\n",
      "Read 32 matches\n",
      "Read 18 matches\n",
      "Read 0 matches\n",
      "Read 0 matches\n",
      "Read 15 matches\n",
      "Read 16 matches\n",
      "Read 24 matches\n",
      "Read 36 matches\n",
      "Read 23 matches\n",
      "Read 21 matches\n",
      "Read 22 matches\n",
      "Read 26 matches\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>f1</th>\n",
       "      <th>f2</th>\n",
       "      <th>f1_start</th>\n",
       "      <th>f1_end</th>\n",
       "      <th>f2_start</th>\n",
       "      <th>f2_end</th>\n",
       "      <th>score</th>\n",
       "      <th>rho</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>22April_2010_Thursday_heute_default-0</td>\n",
       "      <td>29April_2010_Thursday_heute_default-0</td>\n",
       "      <td>10</td>\n",
       "      <td>35</td>\n",
       "      <td>14</td>\n",
       "      <td>39</td>\n",
       "      <td>0.918155</td>\n",
       "      <td>3.789952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>11February_2010_Thursday_tagesschau_default-9</td>\n",
       "      <td>12January_2010_Tuesday_heute_default-9</td>\n",
       "      <td>20</td>\n",
       "      <td>55</td>\n",
       "      <td>114</td>\n",
       "      <td>147</td>\n",
       "      <td>0.928610</td>\n",
       "      <td>25.081400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11February_2010_Thursday_tagesschau_default-9</td>\n",
       "      <td>06April_2010_Tuesday_heute_default-6</td>\n",
       "      <td>17</td>\n",
       "      <td>58</td>\n",
       "      <td>112</td>\n",
       "      <td>146</td>\n",
       "      <td>0.926288</td>\n",
       "      <td>31.536909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11February_2010_Thursday_tagesschau_default-9</td>\n",
       "      <td>20July_2010_Tuesday_heute_default-12</td>\n",
       "      <td>33</td>\n",
       "      <td>59</td>\n",
       "      <td>10</td>\n",
       "      <td>36</td>\n",
       "      <td>0.908951</td>\n",
       "      <td>6.039716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>11February_2010_Thursday_tagesschau_default-9</td>\n",
       "      <td>10January_2010_Sunday_tagesschau_default-12</td>\n",
       "      <td>17</td>\n",
       "      <td>52</td>\n",
       "      <td>83</td>\n",
       "      <td>111</td>\n",
       "      <td>0.915518</td>\n",
       "      <td>6.031500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              f1  \\\n",
       "0          22April_2010_Thursday_heute_default-0   \n",
       "1  11February_2010_Thursday_tagesschau_default-9   \n",
       "2  11February_2010_Thursday_tagesschau_default-9   \n",
       "3  11February_2010_Thursday_tagesschau_default-9   \n",
       "4  11February_2010_Thursday_tagesschau_default-9   \n",
       "\n",
       "                                            f2  f1_start  f1_end  f2_start  \\\n",
       "0        29April_2010_Thursday_heute_default-0        10      35        14   \n",
       "1       12January_2010_Tuesday_heute_default-9        20      55       114   \n",
       "2         06April_2010_Tuesday_heute_default-6        17      58       112   \n",
       "3         20July_2010_Tuesday_heute_default-12        33      59        10   \n",
       "4  10January_2010_Sunday_tagesschau_default-12        17      52        83   \n",
       "\n",
       "   f2_end     score        rho  \n",
       "0      39  0.918155   3.789952  \n",
       "1     147  0.928610  25.081400  \n",
       "2     146  0.926288  31.536909  \n",
       "3      36  0.908951   6.039716  \n",
       "4     111  0.915518   6.031500  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "matches_df = get_matches_all(exp_name)\n",
    "matches_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fadf6d1bd90>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_match_stats(matches_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.91\n",
      "./post_disc Signer03_c3_right_PCA50_10_3_02_7_05_06_13_06_04.lsh64 0.91\n",
      "Read 406 nodes/tokens\n",
      "Read 60 clusters\n"
     ]
    }
   ],
   "source": [
    "\n",
    "results = []\n",
    "#for dtw in np.arange(0.89,0.98,0.01):\n",
    "for dtw in [0.91]:\n",
    "    print(dtw)\n",
    "\n",
    "    results_path = post_disc(exp_name, dtw)\n",
    "\n",
    "    \" get nodes \"\n",
    "    nodes_df = get_nodes_df(exp_name)\n",
    "\n",
    "    \" get clusters \"\n",
    "    clusters_list = get_clusters_list(exp_name)\n",
    "    if len(clusters_list)==0: continue\n",
    "    \n",
    "    scores = evaluate(None, signer_id, matches_df, nodes_df, clusters_list,\n",
    "                      group=3, interp=False, boundary_th=0.5, fast_compute=True, purify=True)\n",
    "\n",
    "    scores['dtw'] = dtw\n",
    "    scores['exp'] = exp_name\n",
    "    \n",
    "    results.append(scores)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('n_match', 313),\n",
       "             ('n_node', 99),\n",
       "             ('n_clus', 27),\n",
       "             ('mean_len', 29.818181818181817),\n",
       "             ('mean_n_type', 2.0707070707070705),\n",
       "             ('ned', 0.54072598494909219),\n",
       "             ('coverage', 0.059437358882040645),\n",
       "             ('matching_P', 0),\n",
       "             ('matching_R', 0),\n",
       "             ('matching_F', 0),\n",
       "             ('purity', 0.55169996494917628),\n",
       "             ('inv_purity', 0.72485103399929895),\n",
       "             ('garbage_ratio', 0.069751139151770072),\n",
       "             ('avg_purity', 0.7474747474747475),\n",
       "             ('clus_purity', 0.51515151515151514),\n",
       "             ('clus_purity_inv', 0.93939393939393945),\n",
       "             ('grouping_P', 0.5555555555555556),\n",
       "             ('grouping_R', 0.859375),\n",
       "             ('grouping_F', 0.6748466257668713),\n",
       "             ('type_P', 0.4),\n",
       "             ('type_R', 0.015384615384615385),\n",
       "             ('type_F', 0.02962962962962963),\n",
       "             ('dtw', 0.91),\n",
       "             ('exp', 'Signer03_c3_right_PCA50_10_3_02_7_05_06_13_06_04')])"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 4),\n",
       "              ('n_clus', 2),\n",
       "              ('mean_len', 14.5),\n",
       "              ('mean_n_type', 1.0),\n",
       "              ('ned', 0.5),\n",
       "              ('coverage', 0.00065893366355756014),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.68518518518518523),\n",
       "              ('inv_purity', 0.81481481481481477),\n",
       "              ('garbage_ratio', 0.18518518518518517),\n",
       "              ('avg_purity', 0.75),\n",
       "              ('clus_purity', 0.75),\n",
       "              ('clus_purity_inv', 1.0),\n",
       "              ('grouping_P', 0.5),\n",
       "              ('grouping_R', 1.0),\n",
       "              ('grouping_F', 0.6666666666666666),\n",
       "              ('type_P', 1.0),\n",
       "              ('type_R', 0.002898550724637681),\n",
       "              ('type_F', 0.005780346820809248),\n",
       "              ('dtw', 0.89000000000000001),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 4),\n",
       "              ('n_clus', 2),\n",
       "              ('mean_len', 21.0),\n",
       "              ('mean_n_type', 1.0),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00095431771963508708),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.51249999999999996),\n",
       "              ('inv_purity', 0.91249999999999998),\n",
       "              ('garbage_ratio', 0.087499999999999994),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 1.0),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.90000000000000002),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 4),\n",
       "              ('n_clus', 2),\n",
       "              ('mean_len', 16.5),\n",
       "              ('mean_n_type', 1.0),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00074982106542756844),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.5161290322580645),\n",
       "              ('inv_purity', 0.64516129032258063),\n",
       "              ('garbage_ratio', 0.16129032258064516),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 0.75),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0.0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.91000000000000003),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 2),\n",
       "              ('n_clus', 1),\n",
       "              ('mean_len', 13.0),\n",
       "              ('mean_n_type', 1.0),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00029538405607752694),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.54166666666666663),\n",
       "              ('inv_purity', 0.875),\n",
       "              ('garbage_ratio', 0.125),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 1.0),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.92000000000000004),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 4),\n",
       "              ('n_clus', 2),\n",
       "              ('mean_len', 16.0),\n",
       "              ('mean_n_type', 0.75),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00064757273832380911),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.6333333333333333),\n",
       "              ('inv_purity', 0.55000000000000004),\n",
       "              ('garbage_ratio', 0.23333333333333334),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 0.5),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.93000000000000005),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 2),\n",
       "              ('n_clus', 1),\n",
       "              ('mean_len', 17.0),\n",
       "              ('mean_n_type', 1.0),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00038627145794753524),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.5),\n",
       "              ('inv_purity', 0.625),\n",
       "              ('garbage_ratio', 0.375),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 0.5),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.94000000000000006),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 4),\n",
       "              ('n_clus', 2),\n",
       "              ('mean_len', 16.75),\n",
       "              ('mean_n_type', 1.25),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00076118199066131946),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.38095238095238093),\n",
       "              ('inv_purity', 0.55555555555555558),\n",
       "              ('garbage_ratio', 0.44444444444444442),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 0.75),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.95000000000000007),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 2),\n",
       "              ('n_clus', 1),\n",
       "              ('mean_len', 25.5),\n",
       "              ('mean_n_type', 1.5),\n",
       "              ('ned', 0.5),\n",
       "              ('coverage', 0.00057940718692130286),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.51020408163265307),\n",
       "              ('inv_purity', 0.59183673469387754),\n",
       "              ('garbage_ratio', 0.40816326530612246),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 1.0),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.96000000000000008),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 2),\n",
       "              ('n_clus', 1),\n",
       "              ('mean_len', 21.5),\n",
       "              ('mean_n_type', 1.0),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00048851978505129456),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.097560975609756101),\n",
       "              ('inv_purity', 0.17073170731707318),\n",
       "              ('garbage_ratio', 0.82926829268292679),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 0.5),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.97000000000000008),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')])]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "resname = 'train_set'\n",
    "final = load_obj(name='grid_post_disc_' + resname, path='/home/korhan/Dropbox/tez/files/ZR_dtw_curves')\n",
    "final['AE (Signer04)2'] = results\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "final = dict()\n",
    "resname = 'train_set'\n",
    "final['AE (Signer04 Signer05 Signer09)'] = results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_obj(obj=final, name='grid_post_disc_' + resname, path='/home/korhan/Dropbox/tez/files/ZR_dtw_curves' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "metadata": {},
   "outputs": [],
   "source": [
    "resname = 'test_set'\n",
    "final = load_obj(name='grid_post_disc_' + resname, path='/home/korhan/Dropbox/tez/files/ZR_dtw_curves')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 475,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 4738),\n",
       "              ('n_clus', 2231),\n",
       "              ('mean_len', 23.992401857323767),\n",
       "              ('mean_n_type', 1.7566483748417054),\n",
       "              ('ned', 0.87164124293785372),\n",
       "              ('coverage', 0.35238360878938013),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.37800400227652425),\n",
       "              ('inv_purity', 0.072995648901209867),\n",
       "              ('garbage_ratio', 0.17223558354293267),\n",
       "              ('avg_purity', 0.5588856057408189),\n",
       "              ('clus_purity', 0.50042211903756861),\n",
       "              ('clus_purity_inv', 0.47952722667792319),\n",
       "              ('grouping_P', 0.05107640354579992),\n",
       "              ('grouping_R', 0.09456819069949199),\n",
       "              ('grouping_F', 0.06632862820337125),\n",
       "              ('type_P', 0.4878048780487805),\n",
       "              ('type_R', 0.075046904315197),\n",
       "              ('type_F', 0.13008130081300814),\n",
       "              ('dtw', 0.88)]),\n",
       " OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 4791),\n",
       "              ('n_clus', 2221),\n",
       "              ('mean_len', 24.179920684616992),\n",
       "              ('mean_n_type', 1.7693592151951576),\n",
       "              ('ned', 0.86950575815739006),\n",
       "              ('coverage', 0.34842668582882647),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.37885732294808877),\n",
       "              ('inv_purity', 0.071234973661699161),\n",
       "              ('garbage_ratio', 0.16528747017243708),\n",
       "              ('avg_purity', 0.5531204341473597),\n",
       "              ('clus_purity', 0.49634731788770614),\n",
       "              ('clus_purity_inv', 0.47568357336672928),\n",
       "              ('grouping_P', 0.052807347109163016),\n",
       "              ('grouping_R', 0.09742010011551791),\n",
       "              ('grouping_F', 0.068489442338928),\n",
       "              ('type_P', 0.5061728395061729),\n",
       "              ('type_R', 0.07692307692307693),\n",
       "              ('type_F', 0.13355048859934854),\n",
       "              ('dtw', 0.89000000000000001)]),\n",
       " OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 4565),\n",
       "              ('n_clus', 2093),\n",
       "              ('mean_len', 24.000657174151151),\n",
       "              ('mean_n_type', 1.7193866374589266),\n",
       "              ('ned', 0.86363760069671203),\n",
       "              ('coverage', 0.32156972126621536),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.38670260385912114),\n",
       "              ('inv_purity', 0.075210956399169512),\n",
       "              ('garbage_ratio', 0.16383169203222919),\n",
       "              ('avg_purity', 0.5504928806133625),\n",
       "              ('clus_purity', 0.49748083242059143),\n",
       "              ('clus_purity_inv', 0.45191675794085434),\n",
       "              ('grouping_P', 0.061555312157721795),\n",
       "              ('grouping_R', 0.10841049382716049),\n",
       "              ('grouping_F', 0.07852452144753388),\n",
       "              ('type_P', 0.5340909090909091),\n",
       "              ('type_R', 0.08818011257035648),\n",
       "              ('type_F', 0.15136876006441224),\n",
       "              ('dtw', 0.90000000000000002)]),\n",
       " OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 3728),\n",
       "              ('n_clus', 1694),\n",
       "              ('mean_len', 23.733369098712448),\n",
       "              ('mean_n_type', 1.6955472103004292),\n",
       "              ('ned', 0.83432009403613039),\n",
       "              ('coverage', 0.25666578419878222),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.41110324483775812),\n",
       "              ('inv_purity', 0.088047197640117994),\n",
       "              ('garbage_ratio', 0.14620648967551622),\n",
       "              ('avg_purity', 0.5592811158798283),\n",
       "              ('clus_purity', 0.50268240343347637),\n",
       "              ('clus_purity_inv', 0.43320815450643779),\n",
       "              ('grouping_P', 0.07591201716738197),\n",
       "              ('grouping_R', 0.1295787545787546),\n",
       "              ('grouping_F', 0.09573748308525035),\n",
       "              ('type_P', 0.59375),\n",
       "              ('type_R', 0.07129455909943715),\n",
       "              ('type_F', 0.1273031825795645),\n",
       "              ('dtw', 0.91000000000000003)]),\n",
       " OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 2977),\n",
       "              ('n_clus', 1116),\n",
       "              ('mean_len', 23.416526704736313),\n",
       "              ('mean_n_type', 1.6231105139402082),\n",
       "              ('ned', 0.92492819234244827),\n",
       "              ('coverage', 0.19078798078741349),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.39287319806994936),\n",
       "              ('inv_purity', 0.1755177270956334),\n",
       "              ('garbage_ratio', 0.13875985254892559),\n",
       "              ('avg_purity', 0.5068861269734632),\n",
       "              ('clus_purity', 0.45649983204568356),\n",
       "              ('clus_purity_inv', 0.45381256298286865),\n",
       "              ('grouping_P', 0.1719852200201545),\n",
       "              ('grouping_R', 0.2788671023965142),\n",
       "              ('grouping_F', 0.21275711614377726),\n",
       "              ('type_P', 0.46808510638297873),\n",
       "              ('type_R', 0.04127579737335835),\n",
       "              ('type_F', 0.07586206896551724),\n",
       "              ('dtw', 0.92000000000000004)]),\n",
       " OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 1506),\n",
       "              ('n_clus', 606),\n",
       "              ('mean_len', 24.157370517928285),\n",
       "              ('mean_n_type', 1.7264276228419655),\n",
       "              ('ned', 0.88221354981515476),\n",
       "              ('coverage', 0.1043124314511554),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.39412186379928316),\n",
       "              ('inv_purity', 0.14709677419354839),\n",
       "              ('garbage_ratio', 0.12338351254480287),\n",
       "              ('avg_purity', 0.5092961487383798),\n",
       "              ('clus_purity', 0.4714475431606906),\n",
       "              ('clus_purity_inv', 0.46945551128818064),\n",
       "              ('grouping_P', 0.12881806108897742),\n",
       "              ('grouping_R', 0.23013048635824437),\n",
       "              ('grouping_F', 0.16517667092379737),\n",
       "              ('type_P', 0.6086956521739131),\n",
       "              ('type_R', 0.02626641651031895),\n",
       "              ('type_F', 0.050359712230215826),\n",
       "              ('dtw', 0.93000000000000005)]),\n",
       " OrderedDict([('n_match', 178351),\n",
       "              ('n_node', 618),\n",
       "              ('n_clus', 282),\n",
       "              ('mean_len', 26.724919093851131),\n",
       "              ('mean_n_type', 2.059870550161812),\n",
       "              ('ned', 0.8918672501823488),\n",
       "              ('coverage', 0.046045913543360691),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.35463580324569127),\n",
       "              ('inv_purity', 0.18065165429613789),\n",
       "              ('garbage_ratio', 0.11234117499056485),\n",
       "              ('avg_purity', 0.5275080906148867),\n",
       "              ('clus_purity', 0.50161812297734631),\n",
       "              ('clus_purity_inv', 0.63754045307443363),\n",
       "              ('grouping_P', 0.07119741100323625),\n",
       "              ('grouping_R', 0.20276497695852536),\n",
       "              ('grouping_F', 0.10538922155688624),\n",
       "              ('type_P', 0.6666666666666666),\n",
       "              ('type_R', 0.0150093808630394),\n",
       "              ('type_F', 0.029357798165137616),\n",
       "              ('dtw', 0.94000000000000006)])]"
      ]
     },
     "execution_count": 475,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 331,
   "metadata": {},
   "outputs": [],
   "source": [
    "results.extend(final['AE (Signer01)'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 333,
   "metadata": {},
   "outputs": [],
   "source": [
    "final['AE (Signer01)'] = results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f47645c1f10>"
      ]
     },
     "execution_count": 238,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4764ed0290>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1,1,squeeze=False)\n",
    "for ky, res in final.items():\n",
    "    plot_curve(axes[0,0], res, {'x': 'coverage', 'y': 'clus_purity'}, label=ky, annot_freq=6)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f475a472e50>"
      ]
     },
     "execution_count": 271,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAEKCAYAAADn+anLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XlYVdX6wPHvYpKZBBzBAQWR+TiA1xxSETU158y0HLBB0xzKbtTNrHtLrWzwWv7MnLBBu85lac5pluI8ATkgIA6oDIqgMq3fHxsPoKIH5XAQ1ud5zoNn7332eTcC71lr7fUuIaVEURRFUe7GzNQBKIqiKBWXShKKoihKiVSSUBRFUUqkkoSiKIpSIpUkFEVRlBKpJKEoiqKUSCUJRVEUpUQqSSiKoiglUklCURRFKZGFqQMoLVdXV9mwYUNTh6EoivJI2bdv32UpZY3Svu6RSxINGzZk7969pg5DURTlkSKESHiQ16nuJkVRFKVEVTZJrF+/Hm9vbzw9PZk+ffod+9PS0ujbty+BgYGEhIRw9OhR/b709HQGDBhA06ZN8fHx4a+//irP0BVFUcpNlUwSeXl5jBkzhnXr1hEdHc2SJUuIjo4udszUqVPR6XQcPnyYxYsXM378eP2+8ePH061bN2JjYzl06BA+Pj7lfQmKoijl4pEbkygLUVFReHp60qhRIwAGDRrEmjVr8PX11R8THR1NREQEAE2bNiU+Pp7k5GSsra3Zvn07ixYtAsDKygorK6tyvwZFKW85OTkkJSVx48YNU4ei3IO1tTXu7u5YWlqWyfmqZJI4e/Ys9erV0z93d3dn9+7dxY4JCgpi5cqVtGvXjqioKBISEkhKSsLc3JwaNWowYsQIDh06RIsWLZg5cyZ2dnblfRmKUq6SkpJwcHCgYcOGCCFMHY5yF1JKUlJSSEpKwsPDo0zOWSW7mwwRERFBeno6Op2OWbNm0axZM8zNzcnNzWX//v2MHj2aAwcOYGdnd9cxDUWpbG7cuIGLi4tKEBWYEAIXF5cybe1VyZaEm5sbZ86c0T9PSkrCzc2t2DGOjo4sXLgQ0LKzh4cHjRo1IisrC3d3d1q1agXAgAEDVJJQqgyVICq+sv4/qlotiZObYPsnBAcHc+LECU6fPk12djZLly6lV69exQ5NT08nOzsbgHnz5tG+fXscHR2pXbs29erV4++//wZg8+bNxcYyFEVRKpOqlSTi/4BtH2FBPl9++SVdu3bFx8eHgQMH4ufnx5w5c5gzZw4AMTEx+Pv74+3tzbp165g5c6b+NLNmzWLIkCEEBgZy8OBB3n77bVNdkaJUKUlJSfTu3RsvLy8aN27M+PHj9R/mtm3bhpOTEzqdDh8fH95///27nuP8+fP07NkTgKysLIYMGUJAQAD+/v60bduWa9euAfD444+XyzVFRkbi5eWFl5cXkZGR+u0jR44kKCiIwMBABgwYoI9r7dq1vPvuu+USG6B1pTxKjxYtWsgHdniZlFMcpTx/+MHPoShVVHR0tEnfPz8/XwYHB8sFCxZIKaXMzc2V4eHhctKkSVJKKbdu3Sp79OghpZTy2rVr0tPTU+7bt++O80yaNEmuXr1aSinl1KlT5cSJE/X7YmNj5Y0bN4x9KVJKKXNycmRKSor08PCQKSkpMjU1VXp4eMjU1FQppZRXrlzRHztx4kQ5bdo0KaX2fdDpdDIzM7PEc9/t/wrYKx/gb27VaknUDtC+Xjh67+MURalwtmzZgrW1NSNGjADA3Nyczz//nAULFpCVlVXsWDs7O1q0aMHJkyfvOM+KFSvo1q0boLUqio5Hent7U61aNQDs7e0ByM/P55VXXqFp06aEhYXRvXt3li9fDmhlgqZMmULz5s0JCAggNjYWgMzMTMLDwwkJCaFZs2asWbMGgEWLFtGrVy86depEaGgov/32G2FhYTg7O1O9enXCwsJYv349oI2LgvZB/vr16/qxBiEEHTp0YO3atWXwXb2/qjVw7dwYLKwhWSUJRXkY7/98jOhzV8v0nL51HZnylF+J+48dO0aLFi2KbXN0dKR+/fp3JIOUlBR27drF5MmTi20/ffo01atX1yeC8PBwunTpwvLlywkNDWXYsGF4eXkVe83KlSuJj48nOjqaixcv4uPjQ3h4uH6/q6sr+/fvZ/bs2cyYMYN58+bx4Ycf0qlTJxYsWEB6ejohISF07twZgP3793P48GGcnZ2ZMWPGHbfjnz17Vv98xIgR/Prrr/j6+vLpp5/qt7ds2ZIdO3YwcODAe35Py0LVakmYW0BNH7hwxNSRKIpiBDt27KBZs2Z06dKFiIgI/PyKJ53z589To0ZhIVSdTkdcXBxvvPEGqampBAcHExMTU+w1f/zxB08//TRmZmbUrl2bjh07Ftvfr18/AFq0aEF8fDwAGzZsYPr06eh0Ojp06MCNGzdITEwE0LccDLFw4ULOnTuHj48PP/74o357zZo1OXfunGHflIdUtVoSALX84e9fQUpQt/MpygO51yd+Y/H19dV389xy9epVEhMT8fT0JCoqinbt2t2zG8bGxuaOOQT29vb069ePfv36YWZmxq+//lqqUju3WiW35lGB1kW0YsUKvL29ix27e/fuYhNv3dzc2LZtm/55UlISHTp0KPYac3NzBg0axMcff6zvartx4wY2NjYGx/gwqlZLArRxiawUuJZs6kgURSmF0NBQsrKyWLx4MaDVYHv99dcZPnw4tra2Bp2jSZMm+k/7ADt37iQtLQ2A7OxsoqOjadCgQbHXtGnThhUrVpCfn09ycnKxP+ol6dq1K7NmzUIbL4YDBw6UeNyGDRtIS0sjLS2NDRs20LVrV6SU+i40KSU//fQTTZs21b/u+PHj+Pv7G3TND6vqJYlaBZ+A1OC1ojxShBCsWrWKZcuW4eXlRZMmTbC2tmbq1KkGn8POzo7GjRvr/wCfOnWKJ554goCAAJo1a0bLli3p379/sdf0798fd3d3fH19ee6552jevDlOTk73fJ/JkyeTk5NDYGAgfn5+d4yN3OLs7MzkyZMJDg4mODiYd999F2dnZ6SUDBs2jICAAAICAjh//nyx2163bt1Kjx49DL7uhyFuZbpHRcuWLeVDLTp0PQ0+agid34O2E8soKkWp/GJiYipFxeNVq1axb98+PvjgA4Nfc+3aNezt7UlJSSEkJISdO3dSu3ZtI0ZZsuTkZAYPHszmzZtLPOZu/1dCiH1SypalfT+jjkkIIboBMwFzYJ6Ucvpt+52A74D6BbHMkFIuNGZM2FQHp3qqJaEoVVTfvn1JSUkp1Wt69uypr8IwefJkkyUIgMTExGJ3Ohmb0ZKEEMIc+AoIA5KAPUKIn6SURRduGANESymfEkLUAP4WQnwvpcw2VlyANnitboNVlCrrhRdeKNXxhoxDlJfg4OByfT9jjkmEACellHEFf/SXAr1vO0YCDkKbJWIPpAK5RoxJU9sfLp+AHFUXX1EU5V6MmSTcgDNFnicVbCvqS8AHOAccAcZLKfONGJOmlj/IPLgUc/9jFUVRqjBT393UFTgI1AV0wJdCCMfbDxJCvCSE2CuE2Hvp0qWHf1dVnkNRFMUgxkwSZ4F6RZ67F2wragSwsqD+1EngNND0tmOQUs6VUraUUrYsOlvygVX3AEtbk49L5Fy4wI2/j5Nz/jz5mZk8aneaKYpS+Rnz7qY9gJcQwgMtOQwCBt92TCIQCuwQQtQCvIE4I8akMTODmr4ma0nciInh8ty5ZPy2AfKL9K6Zm2Pu4ICZo2PBVwfMHRwxd3LEzMERc0cHzBwcMHd01H/V/q3tE9bWalEYpdJbvXo1ffv2JSYmRj/BLD4+Hh8fn2IznF977TWGDh16x+sHDBjAxx9/TKNGjViwYAGff/45Qgjy8/P58MMP6d27N++++y7t27fX11syltOnTzNo0CBSUlJo0aIF3377LVZWVqxZs4bJkydjZmaGhYUFX3zxBW3btiU7O5vOnTuzZcsWLCzKp2CG0d5FSpkrhBgL/IZ2C+wCKeUxIcSogv1zgP8Ai4QQRwABvCmlvGysmIqp7Q/HVpdreY6svXu5PHcumdt3YGZvj8vIkVj7+ZGXcZX8qxkFX6+Sp/93BjcvndL2Xb2KvN+ShJaWmDs4FCYaR8fCROPoUCTROGLd1Jtqnp7lct2KUpaWLFlC27ZtWbJkSbE1Ixo3bszBgwfv+dpjx46Rl5dHo0aNSEpK4sMPP2T//v04OTlx7do1bnVn//vf/zbqNYA2Y/zNN99k4sSJDBo0iFGjRjF//nxGjx5NaGgovXr1QgjB4cOHGThwILGxsVhZWREaGsqPP/7IkCFDjB4jGHmehJTyV+DX27bNKfLvc0AXY8ZQolr+sG8RXD0LTu5GexspJdd+/52Uud9wff9+zJ2dqTFhAtUHP4u54x3DL/eUn51NfoaWMLSvGeRfvVIsqei/Xr1KXsZVcs6dIy8jg/wrV5A5OYUnMzPDdewYXF9+GWFuXsZXrSjGce3aNf744w+2bt3KU089VeLCQiX5/vvv6d1bu8ny4sWLODg46EuC29vb6/89fPhwevbsyYABA/j111957bXXsLOzo02bNsTFxbF27Vree+89EhMTiYuLIzExkQkTJjBu3DgAvvvuO/773/+SnZ1Nq1atmD17Nubm5tjb2/Pyyy+zadMmvvzyS7Zs2cIPP/wAwLBhw3jvvfcYPXq0Pg7Qyo4X7SHo06cPb731VuVIEhVa0cHruyWJaxchaS/cvAr+/cHcslSnl3l5XF2/npRv5nEzNhaLunWo9c47PNa/H2YPWJjLzMoKMxcXLFxcHuj1+Tdvai2V9HQufz2Xy/+dRdZfu6j7ycdYmnBykPIIWhdR9tWUawfAk/deL37NmjV069aNJk2a4OLiwr59+/Tlw0+dOoVOp9MfO2vWLNq1a1fs9Tt37uTZZ58FICgoiFq1auHh4UFoaCj9+vXjqaeeKnb8jRs3ePnll9m+fTseHh76194SGxvL1q1bycjIwNvbm9GjR3Py5El+/PFHdu7ciaWlJa+88grff/89Q4cOJTMzk1atWvHpp59y+fJlHnvsMX230e1lwletWsVbb73FxYsX+eWXX/Tb/f392bNnj6Hf1YdWdZPErRpOyUegcUftBz5pT+EjPbHw2N1fQ/954NL4vqfNz87myqrVpMyfT05iIlaNG1Nn2jScevZAWJYu0ZQ1s2rVMKtRA4saNaj7ycfYtWnDhf/8h9O9+1Bn2lQcOnUyaXyKcj9Llixh/PjxAAwaNIglS5bok4Qh3U1FS4Wbm5uzfv169uzZw+bNm5k4cSL79u3jvffe0x8fGxtLo0aN8PDwAODZZ59l7ty5+v09evSgWrVqVKtWjZo1a5KcnMzmzZvZt2+fftLb9evXqVmzpv49b68NVZK+ffvSt29ftm/fzuTJk9m0aZP+HFZWVmRkZODg4GDQuR5G1U0S1RygekPYOQt+/xjyCiZ5O7qDe0sIeQncg+HqOVg7Eea0g+4fg27IXccw8jMzSfvxf6QuWkTuxYtYBwRQc9Z/cQgNRZiZ+k7jOwkheKxvH2yCgjg76XWSXhlD9SFDqPnPNzArKH2sKCW6zyd+Y0hNTWXLli0cOXIEIQR5eXkIIfjkk08MPsftpcKFEISEhBASEkJYWBgjRowoliTup1qR35VbpcJvFeebNm3aHcdbW1tjXtC96+LiQnp6Orm5uVhYWJCUlFRslbxb2rdvT1xcHJcvX8bV1RWAmzdvYm1tbXCcD6Pi/fUqT82HaU3cVqNg4LfwWgy8dgwGRsLjr0L9f4B/Pxj9J7g1hzVjYNlwrUhggdy0NC79dxYnOoVy8eOPsWrUiPoL5tPwfz/iGBZWIRNEUdUaedBw6VKchw0l7fvviR/4DDdPnTJ1WIpyh+XLl/P888+TkJBAfHw8Z86cwcPDgx07dhh8Dh8fH30F2HPnzrF//379voMHD95RJtzb25u4uDh9efGiC/+UJDQ0lOXLl3Px4kVAS24JCQl3HCeEoGPHjvo1MiIjI/XjJSdPntTfEr9//35u3ryJS0E3c0pKCq6urliWU89E1W1JALR7TXvcj5MbDF0Df86CLf+BpD3ktJtO6qZo0pYtQ2ZlYd85FNeXXsImMND4cZcxMysrar31FratW3P+rbc5PeBpav/rbZz691e31CoVxpIlS3jzzTeLbevfv79+++1jEuHh4fqB5Ft69OjBtm3b6Ny5Mzk5OUyaNIlz585hbW1NjRo1mDNnTrHjbWxsmD17Nt26dcPOzs6gukm+vr588MEHdOnShfz8fCwtLfnqq6/uSEAAH330EYMGDeKdd96hWbNmjBw5EtDW4V68eDGWlpbY2Njw448/6n8Xy7NMOFTFUuEPKTvqVy5P+ydX/s4FzHDq0QOXl16i2m3r4j6qcpIvcu7NN8natQuHJ7tR5/33S30XllI5VYZS4devX6djx47s3LlT3+1zP7fKhEspGTNmDF5eXkycaLplBvr168f06dNp0qRJiceUZanwit0XUoHciIkhaeJETg1/g6txFlRvVY/G3S9Q1zOKatUrz6dty1o1qT9/HjVee42MDRs53acvWSWsqqUojxobGxvef//9YncR3c8333yDTqfDz8+PK1eu8PLLLxsxwnvLzs6mT58+90wQZU21JO4ja+9eLn89l8wd2gS46s8+i/OwoVi4ukLMz/DTq5B7E7pNh+ZDK9W62dcPHuTs65PIuXCBGq++isuLL6g5FVVYZWhJVBWqJVEObvx9nPjBQ0h47nluHDtGjQkT8NyymZqvv6YlCACfp7RBbfdg+Hkc/O95yEo1beBlyEanw2P1Khy7duHSF1+QGD6SnOSLpg5LUZRypJJECcwd7MlNuUytd97Bc/MmXEe9fPe+ece68PxqCPsP/L0e/q8NxP1e/gEbibmDA3U//ZQ6H37A9cOHOd27Nxlbt5o6LEVRyolKEiWwrFuXxuvX4/zckPvPkDYzgzbj4IVNYGUHi3vDhsmQa9wF9sqLEILH+vfHY8VyLGrXJmn0K1z4cCr52ZXj+hRFKZlKEvdQ6ts/6+rg5e3QYjj8+V+Y3xkuHTdKbKZQrVEjGv64lOrPP0/at98S/8wgbsadNnVYiqIYkUoSZc3KFp76Agb9AOln4Ov2sHehVm22EjCrVo3a/3ob99mzyT1/ntP9+5O+YqVaC0MpF0lJSfTu3RsvLy8aN27M+PHjyS5o0W7btg0nJyd0Oh0+Pj4lFv87f/48PXv2BCArK4shQ4YQEBCAv78/bdu25dq1awA8/vjj5XJNkZGReHl54eXlRWRkpH57u3bt0Ol06HQ66tatS58+fQBYu3Yt7777brnEBmhVSh+lR4sWLeQj48o5KSN7SznFUcolg6W8dtnUEZWp7AsXZPzzQ2W0d1OZ9NrrMvfqVVOHpBhRdHS0Sd8/Pz9fBgcHywULFkgppczNzZXh4eFy0qRJUkopt27dKnv06CGllPLatWvS09NT7tu3747zTJo0Sa5evVpKKeXUqVPlxIkT9ftiY2PljRs3jH0pUkopc3JyZEpKivTw8JApKSkyNTVVenh4yNTU1DuO7devn4yMjJRSat8HnU4nMzMzSzz33f6vgL3yAf7mqpaEMTnWgedWQpcP4cQG+L/H4VTlGfS1rFWL+gsXUGPCeK6uX8/pvv24fuiQqcNSKqktW7ZgbW3NiBEjAK1W0ueff86CBQvIysoqdqydnR0tWrTQl+AoasWKFXTr1g3QWhVF6yV5e3vr6zHdKtedn5/PK6+8QtOmTQkLC6N79+76UhoNGzZkypQpNG/enICAAGJjYwGtvHd4eDghISE0a9aMNWvWALBo0SJ69epFp06dCA0N5bfffiMsLAxnZ2eqV69OWFgY69evLxbv1atX2bJli74lIYSgQ4cOrF279uG+oQaq2mU5yoOZGTw+Fjzaw4oX4Ns+0HoshL4LFo9+IT1hbo7rqFHYhrTi3KRJxA95jhrjx+EycmSFr1ulPLiPoj4iNjW2TM/Z1Lkpb4a8WeL+Y8eO6Su+3uLo6Ej9+vXvSAYpKSns2rWLyZMnF9t++vRpqlevrk8E4eHhdOnSheXLlxMaGsqwYcPwuq16wsqVK4mPjyc6OpqLFy/i4+NDeHi4fr+rqyv79+9n9uzZzJgxg3nz5vHhhx/SqVMnFixYQHp6OiEhIfpV7vbv38/hw4dxdnZmxowZ1KtXuMrz7eXCQVuJLzQ0FMcid1e2bNmSHTt2MHDgwBK/X2VF/RaXlzqB8NI2aDkS/voSFnTT1qyoJGybN8Nj9SocOnfm0qefceaFF8i5WHmuT3k07Nixg2bNmtGlSxciIiLw8/Mrtr9oqXAAnU5HXFwcb7zxBqmpqQQHBxMTE1PsNX/88QdPP/00ZmZm1K5dm44dOxbb369fPwBatGihLwS4YcMGpk+fjk6no0OHDty4cYPERG35gVstB0MtWbLkjnUsatasyblz5ww+x8NQLYnyZGULPT/T1q9Y8SLM66x1R7lWjmVEzR0dcfv8M9Ifb03y1Gmc7tOXutOnYd++valDU8rYvT7xG4uvr6++m+eWq1evkpiYiKenJ1FRUbRr1+6e3TC3lwoHrVupX79+9OvXDzMzM3799ddSzSy/1Sq5VSoctLHeFStWFFtzG2D37t3Y2dnpn7u5ubFt2zb986SkJDp06KB/fvnyZaKioli1alWx89y4cQObB1y8rLRUS8IUfJ6C4b9AdibMD4PE3aaOqMwIIag+cCAey5dh4erKmZdeJnnadDWnQnlooaGhZGVlsXjxYkBbI/r1119n+PDh2NraGnSOJk2a6D/tg7ZSXVqaVvo/Ozub6OjoO6q1tmnThhUrVpCfn09ycnKxP+ol6dq1K7NmzdLf9XeghPpnXbt2ZcOGDaSlpZGWlsaGDRvo2rWrfv/y5cvp2bPnHWtHHD9+HH9/f0Mu+aGpJGEq7i1g5AaweQwW94Lon0wdUZmq5ulJw2X/o/rgwaRGRhLfvz/pq1arZKE8MCEEq1atYtmyZXh5edGkSROsra2ZOnWqweews7OjcePG+jGMU6dO8cQTTxAQEECzZs1o2bLlHSvH9e/fH3d3d3x9fXnuuedo3rw5Tk5O93yfyZMnk5OTQ2BgIH5+fneMjdzi7OzM5MmTCQ4OJjg4mHfffbdYV9TSpUvv6GqC8i0Xrgr8mVrmZVgySFtP+8mPoJXpKkwaS8aWLVz87DOyT57C3MWF6oMGUX3QM1gU6RtWKr7KUuBv1apV7Nu3jw8++MDg19wqF56SkkJISAg7d+6ktonWhU9OTmbw4MFs3ry5xGPKssCfGpMwNTtXGPqTdufTun/ClTPQ+d/aXVGVhEOnTth37Ejmn3+StvhbLn/1FSlz5+LYvTvVhz6PzW2Di4piTH379iUlJaVUr+nZsyfp6elkZ2czefJkkyUIgMTERD799NNyez/Vkqgo8vO0JLFnHvj1g75zKsUtsndz8/Rp0r77nvRVq5BZWdi0aIHz0KE4hHZCWKjPLRVVZWlJVAWqVHhlZGYO3WdA5/fh2Er4tm+xtbQrk2oeHtSe/A5ev2+jZsSb5F64wNnx4znZpQsp8+eTd+WKqUNUFKWAShIViRDQdgL0mwdnomB+V0hPNHVURmPu4IDL8OE03vAb7l/Owsq9Hhc/mcGJDh05//773IyLe+Bzr1+/Hm9vbzw9PZk+ffod+9PS0ujbty+BgYGEhIRw9OjRh7kURam0VJKoiAKfhudXQsYFmBcG5w+bOiKjEubmOHTuTIPFkdoiR08+yZUVK4nr3oPEF17k2vbtyPx8g8+Xl5fHmDFjWLduHdHR0SxZsoTo6Ohix0ydOhWdTsfhw4dZvHgx48ePL+vLUpRKQSWJisqjPYSv17qhFj4JJ0u+k6EysW7alLpTP8Rz21ZqjB/Hzb//5sxLLxPXoyepP/xAfmbmfc8RFRWFp6cnjRo1wsrKikGDBulr59wSHR1Np06dAGjatCnx8fEkJycb5ZoU5VGmkkRFVstXW8ioekP4YSAc+N7UEZUbC2dnXEePxnPzJup+8jFmdnYk//s/nOjQkeSPPyHnHgvZnz179r71cIKCgli5ciWgJZWEhASSkpKMczFKmTE3N0en0+Hn50dQUBCffvop+QWtzG3btulLgN8yfPjwO2Zp3zJhwgS2b98OaOW3mzVrRlBQEL6+vnz99dcAzJkzRz95z5hSU1MJCwvDy8uLsLAw/QS/qKgofbnwoKCgYjOvO3furD/OmFSSqOgc68KIX6FBG1jzCvz+caVZm8IQwsoKp6eeouGy/9Hghx+wa9uG1MhIToZ1IenVcWTt3ftAa1lERESQnp6OTqdj1qxZNGvWDHNzcyNcgVKWbGxsOHjwIMeOHWPjxo2sW7euxHUj7uVWAcD27duTk5PDSy+9xM8//8yhQ4c4cOCAvjTGqFGjGDp0aBlfRXF5eXlMnz6d0NBQTpw4QWhoqH4czd/fn71793Lw4EHWr1/Pyy+/rC/98fzzzzN79myjxgZGThJCiG5CiL+FECeFEBF32f+GEOJgweOoECJPCGF45auqwtoJhiyHwEGw9UP4eRzk5Zo6qnIlhMC2eTPcP/8cz00bcRk5ksyoKBKee15b+KjIbG43NzfOnDmjf21SUlKxctCgVQ9duHAhBw8eZPHixVy6dIlGjRqV6zUpD6dmzZrMnTuXL7/8stQfFIqWC8/IyCA3NxcXFxdAq8V0q+bSe++9x4wZMwDYs2cPgYGB6HQ63njjDX1ZjEWLFtGvXz+6deuGl5cX//znP/Xvs2HDBlq3bk3z5s15+umn9QsaNWzYkDfffJPmzZuzbNky1qxZw7BhwwAYNmwYq1evBsDW1haLgtvCb9y4UWy1zF69erFkyZLSfdMegNFuShdCmANfAWFAErBHCPGTlFI/giil/AT4pOD4p4CJUspUY8X0SLOw0uZOOLnDjhnaoPaAhVDN3tSRlTvLOnWo+fpruL4ymis//Uzqt4s5/9ZbXJwxg+pDBtN88GBOnDjB6dOncXNzY+nSpfzwww/FzpGeno6trS1WVlbMmzeP9u3bFyvFDNodUuPHjycvL48XXniBiIjin3OuXLnSCFVdAAAgAElEQVTCc889R2JiIrm5uUyaNEm/1gFonxBbtmyJm5tbudX+Ly8Xpk7lZkzZlgqv5tOU2m+/XarXNGrUiLy8PC4WVBzesWMHOp1Ovz8xMfGOLijQajYNGDAA0Epj9OrViwYNGhAaGkrPnj159tlnMbttQuuIESP45ptvaN269R0/CwcPHuTAgQP6BPPqq69iY2PDBx98wKZNm7Czs+Ojjz7is88+068q5+Liwv79+wGtxVKnTh0AateuXWx8bPfu3YSHh5OQkMC3336rTxrVq1fn5s2bpKSk6BOcMRizJRECnJRSxkkps4GlQO97HP8sYPy0+CgTAkInQ8/P4eQmWNSjUpUbLy0zGxuqPzOQRj//TL3587Dx9+fyf2eR8GR3pvXpQ9cuXfDx8WHgwIH4+fkxZ84c5syZA2iTjfz9/fH29mbdunXMnDmz2LkNuUPqq6++wtfXl0OHDrFt2zZef/11/VKaADNnzlSTz8pZu3btOHjwoP7Rq1evux53e8nwefPmsXnzZkJCQpgxY0ax9SJA+1CRkZFB69atARg8eHCx/aGhoTg5OWFtbY2vry8JCQns2rWL6Oho2rRpg06nIzIykoSEBP1rnnnmmbvGJoQo1mJo1aoVx44dY8+ePUybNq1YFdvyKBluzOmtbsCZIs+TgFZ3O1AIYQt0A8aWsP8l4CWA+vXrl22Uj6KW4eBQF5aPKCg3vgJcve7/ukpKCIF9mzbYt2nD9SNHufT55wT+uo5f6tWjxthXceqt/aEYNWqU/jWtW7fm+PHjJZ6z6B1SgP4OKV9f32Lvm5GRgZSSa9eu4ezsrP+Ul5SUxC+//MK//vUvPvvsM2NctkmV9hO/scTFxWFubk7NmjXvWAfiXu5WMjwgIICAgACef/55PDw8WLRokcHnu1UuHApLhkspCQsLK7FLqGjJ8Fq1anH+/Hnq1KnD+fPnqVmz5h3H+/j4YG9vz9GjR2nZUps4XR4lwyvKwPVTwM6SupqklHOllC2llC1rqKJwGu9uMHxtkXLju0wdUYVgE+BP/QXzqb9wARbOLpx/+23ievcmY/PmUvVbG3KH1NixY4mJiaFu3boEBAQwc+ZMfRfFhAkT+Pjjj+/oslDKzqVLlxg1ahRjx44t9snbED4+PvpKsNeuXStW/vvgwYN3lAt/7LHHcHBwYPduraz/0qVL7/se//jHP9i5c6f+fTIzM0v8YNKrVy8iIyMBiIyMpHdvrdPl9OnT+oHqhIQEYmNjadiwIaCtWXHhwgX9c2Mx5k/wWaBekefuBdvuZhCqq6n03FrACxvBxhkie0H0mvu/poqwa92ahsv+h9vMmZCXT9KYsSQMepbMqKgye4/ffvsNnU7HuXPnOHjwIGPHjuXq1ausXbuWmjVr3rHUpvLwrl+/rr8FtnPnznTp0oUpU6aU+jw9evTQJwYpJR9//DHe3t7odDqmTJly11bE/PnzefHFF9HpdGRmZt63XHiNGjVYtGgRzz77LIGBgbRu3Vq/BvbtIiIi2LhxI15eXmzatEk/5vHHH38QFBSETqejb9++zJ49G1dXVwD27dvHP/7xD33r1WiklEZ5oHVlxQEegBVwCPC7y3FOQCpgZ8h5W7RoIZXbXLss5TehUk5xkvKv/zN1NBVOfk6OTFu2TB5v/4SM9m4qE154UV4/duyer/nzzz9lly5d9M+nTp0qp06dWuyY7t27y+3bt+ufd+zYUe7evVtGRERINzc32aBBA1mrVi1pY2MjhwwZUrYXZQLR0dGmDqFMtWnTRqalpRl8fEZGhv7f06ZNk+PGjTNGWAYbN26c3LRp01333e3/CtgrH+Rv+YO8yOCTQ3fgOHAK+FfBtlHAqCLHDAeWGnpOlSRKcDNTyiWDpZziKOX6t6XMyzN1RBVO3vXr8vK8+TI2pJWM9m4qkya+Jm/Gx9/12JycHOnh4SHj4uLkzZs3ZWBgoDx69GixY0aNGiWnTJkipZTywoULsm7duvLSpUvFjtm6davs0aOHUa6nvFW2JLFr1y556NAhg49funSpDAoKkn5+frJ79+7y4sWLRozu/ubOnVvivrJMEqpUeGWSnwfrIyBqLvj1hT5zwNL6/q+rYvKuXiVlwQJSIxcjc3Kw8ffH0t0dSzc30v/3P1xeeAGXkeH8+uuvTJgwgby8PMLDw/nXv/6lvztq1KhRnDt3juHDh3P+/HmklERERPDcc88Ve69t27YxY8aMSnELrCoV/ugoy1LhKklUNlLCn7Ng42So/zgM+h5s1fzEu8m9dImUhYu4ER1NTlISOefPQ14eLqNepuaECaYOr8KJiYmhadOmpR4kVsqXlJLY2Fi1Mp1SAiGgzTitnMfq0bCgGzy3HB5Ttw7fzqJGDWr98w39c5mbS+7ly2pZ1RJYW1vrJ26pRFExSSlJSUnB2rrsehBUkqisAgaAfS1YOkSbSzFkGdQJMnVUFZqwsMDShMtSVnTu7u4kJSVx6dIlU4ei3IO1tTXu7u5ldj7V3VTZXYyB7wbAjXQYGAmenU0dkaIoJqCWL1XurqZPQblxD/h+IBz4ztQRKYryCFFJoipwrKOVG/doB2vGwLaPqlS5cUVRHpxKElWFtSMMXgZBz8K2qfDTq5CXY+qoFEWp4NTAdVViYQV9/k8rN779E8g4D09HVsly44qiGEa1JKoaIaDTO/DUTDi1FRZ1hwy1trOiKHenkkRV1WI4PLsELp+A+Z3hUsllsxVFqbrumySEEE8JIVQyqYyadIXhv0DOda3ceMJfpo5IUZQKxpA//s8AJ4QQHwshmho7IKWcuTWHkRvBzhUW94Zjq00dkaIoFch9k4SU8jmgGVol10VCiL+EEC8JIRyMHp1SPpw9tERRVwfLhsNfs00dkaIoFYRB3UhSyqvAcrR1qusAfYH9QohXjRibUp5snWHoGmjaA357C9a/Dfn5po5KURQTM2RMorcQYhWwDbAEQqSUTwJBwOvGDU8pV5Y2MHAxhLwMu77S1tDOuXH/1ymKUmkZMk+iH/C5lHJ70Y1SyiwhxEjjhKWYjJk5PPkRPFYPNrwD15Jh0A+q3LiiVFGGdDdduD1BCCE+ApBSbjZKVIppCQGPvwoDFsDZfbCgK6QlmDoqRVFMwJAkEXaXbU+WdSBKBeTfH55frbUm5ofBuYOmjkhRlHJWYpIQQowWQhwBmgohDhd5nAYOl1+Iikk1bAPhG8DcSmtRrH9bzdBWlCqkxPUkhBBOQHVgGhBRZFeGlDK1HGK7K7WehIlkXIBN78PhH8HcUpux3WaCVmFWUZQKr8zXuBZCOEoprwoh7jpiaapEoZKEiaWcgh2fwaElYGYBLYZpycLJzdSRKYpyD8ZIEmullD0LupckUHRRWymlbPRgoT4clSQqiNTT8MdncPAHEGbQ7HloO1G7K0pRlAqnzJNEwUkFUE9KmfgwwZUllSQqmLQE+OPzwhXvmj2nJYvqDUwbl6IoxRhl+VKpZZBfHjgqpfKr3gCe+gLGHYDmQ+Hg9zCrubaoUeppU0enKMpDMuQW2P1CiGCjR6I82h6rBz0/g3EHoWU4HPoRZrWA1WMgNc7U0SmK8oDu2d0EIISIBTyBBCATbWxCSikDjR/enVR30yPi6nnYORP2LdSWSQ18BtpPApfGpo5MUaoko4xJFJz4rp3LUkqTTMFVSeIRk3EBdv4X9i6AvJsQ8DS0mwQ1mpg6MkWpUowyJlFAlvBQlPtzqA3dpsKEw9B6DMT8DF+FwPKRcOlvU0enKMp9GJIkfgHWFnzdDMQB6ww5uRCimxDibyHESSFERAnHdBBCHBRCHBNC/G5o4Mojxr4mdPkAxh+GNuPg73XwVStYNgKSo00dnaIoJbhvd9MdLxCiOfCKlPKF+xxnDhxHq/2UBOwBnpVSRhc55jHgT6CblDJRCFFTSnnxXudV3U2VRGYK/PUlRM2F7Gvg2xva/xNq+5s6MkWplIzZ3VSMlHI/0MqAQ0OAk1LKOCllNtqCRb1vO2YwsPLWPIz7JQilErFzgc5TYMIRaP8GnNoKc9rA0iFwXpUGU5SK4r7rSQghXivy1AxoDpwz4NxuwJkiz5O4M7k0ASyFENsAB2CmlHKxAedWKgtbZ+j0jjZesev/YNcciF0L3j3giTegbjNTR6goVZohLQmHIo9qaGMTt7cIHpQF0ALoAXQFJgsh7rjtpWBN7b1CiL2XLl0qo7dWKhSb6tDxbW2Au8PbkPAHzO0APzyjrWmhKIpJ3LMlIYSogZYUTkop00t57rNA0UI+7gXbikoCUqSUmUCmEGI72rKox4seJKWcC8wFbUyilHEojxKbx6DDm/CP0RD1Nfz1FXzTCTzDoEMEuJe6S1VRlIdwr/UkXgCOAbOAWCFEr1Keew/gJYTwEEJYAYOAn247Zg3QVghhIYSwReuOiinl+yiVkbWjNlYx4QiEvqu1JuaFwrf9IHG3qaNTlCrjXt1NEwA/KWVr4HHgrdKcWEqZC4wFfkP7w/8/KeUxIcQoIcSogmNigPVoixhFAfOklEdLfxlKpVXNAdq9riWLzu/D+UOwoAss7g0Jf5k6OkWp9O5VKny/lLJ5Sc9NRd0CW8VlZ2qzt3fOhMxL0LCd1g3VsK2pI1OUCs0Y60lcRLtt9ZZBRZ9LKceV9s3KgkoSCgDZWVpdqJ0ztTW4G7TVxjIatgMh7v96RaliHjRJ3Gvg+o3bnqtbTJSKw8pWu222ZTjsi4SdX0DkU1C/NTzxJjTqoJKFopSBUs+4NjXVklDuKucGHPhWW1o14xzUawVP/BMah6pkoSiU44xrRamQLK0h5EUYfxB6fApXzsJ3/WFeZzixER6xD0OKUlGoJKFULhbVIPgFGLcfen4B1y7C9wO0iXk7Z2rFBFXCUBSD3W+Na3NgnJTy8/IL6d5Ud5NSKrnZcHgp7P4akgvurnZ0B68w7eHxBFSzN22MilIOjLnoUJSUMuSBIytjKkkoD+zKWTi5Uet+itumVZ81t4IGj2szur26gKuXGsNQKiVjJonPAUvgR7TlSwF9Ndhyp5KEUiZysyHxr8KkcSlW2/5Yg4JWRhftdlorW9PGqShlxJhJYutdNkspZafSvllZUElCMYr0RC1ZnNgIp3+HnCwwrwYe7QpaGWFqfW7lkWa0JFHRqCShGF3ODUjYCSc3wYkNkHJS2+7cuHAso0Fb7Y4qRXlEGLMlUQuYCtSVUj4phPAFWksp5z9YqA9HJQml3KXGwYmChBG/A3JvgIUNeLQvTBrVG5o6SkW5J2MmiXXAQuBfUsogIYQFcEBKGfBgoT4clSQUk8q5DvF/aAnjxAZIi9e2uzbRxjG8wrRZ3xbVTBqmotzOmElij5QyWAhxQErZrGDbQSml7gFjfSgqSSgVhpSQckpLFic3askjLxss7bSyIF6dtfGMx+rd70yKYnTGqN10S6YQwgWQBW/0D+BKad9IUSodIcDVU3u0fkWrUHt6e+EA+N+/aMfV9NVaGJ5hUP8fYG5p2rgVpRQMaUk0R1t4yB84CtQAnpZSHjJ+eHdSLQnlkSAlXPq74BbbDdraF/k5UM2xoJVRkDQc65g6UqWKMGZ3UzUgD/AGBPA3YCalvPkggT4slSSUR9LNDIj7vWAsY6NWhBCgdkDhRD73YDA3pHGvKKVnzCRxx2JDplyASCUJ5ZEnJVyMLkwYibtA5oG1EzTupCUMz85gX9PUkSqVSJmPSQghagNugI0QohlaKwLAEVDTUBXlQQkBtfy0R9uJcD1dKxNyYqPWPXVslXZcHV3BHVNdwK05mJmXaRjr169n/Pjx5OXl8cILLxAREVFsf1paGuHh4Zw6dQpra2sWLFiAv79/mcagVHz3WpluGDAcaAnsoTBJZACLpJQryyPA26mWhFKp5edD8pGCVsYmSIoCmQ82zuAZqiWMxqFg5/JQb5OXl0eTJk3YuHEj7u7uBAcHs2TJEnx9ffXHvPHGG9jb2zNlyhRiY2MZM2YMmzdvftgrVEykzFsSUspIIFII0V9KueKholMUxTBmZlAnSHu0fwOyUuHUloLZ3xvhyDJAgFuLglZGZ6jTTHtdKURFReHp6UmjRo0AGDRoEGvWrCmWJKKjo/Wti6ZNmxIfH09ycjK1atUqs8tVKj5DfrLchRCOQjNPCLFfCNHF6JEpigK2zhAwAPrOgUkn4MUt0OEtbd+2afBNJ5jhBatGwZHlWlIxwNmzZ6lXr3D+hru7O2fPni12TFBQECtXah0GUVFRJCQkkJSUVDbXpTwyDLmVIlxKOVMI0RVwAZ4HvgU2GDUyRVGKMzPTWhBuLaDDm5B5WWtlnNgAx3+DQ0tAmIF7iNbC8OoCtQMfuPR5REQE48ePR6fTERAQQLNmzTA3L9txEaXiMyRJ3PoJ6w4sllIeE0IV3FcUk7NzhcCB2iM/D87uLywXsuUD7WFfW7tTyisMGnfU7qAC3NzcOHPmjP5USUlJuLm5FTu9o6MjCxcuBEBKiYeHh757Sqk6DEkS+4QQGwAP4C0hhAOQb9ywFEUpFTNzqBesPTr9S1u29VYV29if4eB3IMy1Gd9eYQR7dOLEiROcPn0aNzc3li5dyg8//FDslOnp6dja2mJlZcW8efNo3749jo6OJrpAxVQMSRIjAR0QJ6XMKijRMcK4YSmK8lDsa4JusPbIy4WkPdrttcc3wKb3sOA9vgx1omvb5uSZWxM+8kX8/PyYM2cOAKNGjSImJoZhw4YhhMDPz4/5801S+FkxMUMm07W/23Yp5XajRHQf6hZYRXlIV88XtjJObYXsDDCzhAatCybyhUENb7WMayVjzBnXPxd5ag2EAPvUynSKUgnk5Wgzvm8t43oxWtvuVL9w8NujPVjZmTZO5aGV28p0Qoh6wBdSyv6lfbOyoJKEohjRlaTCKrZx2yAnE8ytoEGbwtnfLo1VK+MRVJ5JQgDHpJS+9z3YCFSSUJRyknsTEv8qTBqX/9a2V29YmDAatgVLG5OGqRjGmN1NsyhYSwJt8p0OiJdSPmdAUN2AmYA5ME9KOf22/R2ANcDpgk0rpZT/vtc5VZJQFBNJiy+oL7VJq2ibex0srKFhu8JlXJ3VLbIVlTGTxLAiT3PREsROAwIyB44DYUASWv2nZ6WU0UWO6QBMklL2NDRglSQUpQLIuQEJOwtaGRsg9ZS23cWzoPR5mNZFZWlt2jgVPaOtTFdQw+lBhAAnpZRxAEKIpUBvIPqerzKSpIwk/jz3Jz0a9cDOUg3CKcpDsbTWCg56hsKT07VlXG/dMbV3Aez+P7C01Qa9by2wVL2BqaNWHsC9SoUfobCbqdguQEopA+9zbjfgTJHnSUCruxz3uBDiMHAWrVVx7C6xvAS8BFC/fv37vO3dxaTG8J9d/yGwRiBNnZs+0DkURSmBS2Pt0eplyM7S1vu+Nfv7+HrtGFfvwm6p+o+DhZVpY1YMcq+WhMFdQA9hP1BfSnlNCNEdWA143X6QlHIuMBe07qYHeaPatrUBuJB5QSUJRTEmK1to0kV7SAkpJwsXWIqaC399CVb22jKut0qGOLmbOmqlBPdKEpZArdvHH4QQbYALBpz7LFCvyHP3gm16UsqrRf79qxBithDCVUp52YDzl0otO628cXJmclmfWlGUkggBrl7ao/UYuHkNTm8vnJcRu1Y7rqZf4byMeq3A3NK0cSt690oSXwBv3WX71YJ9T93n3HsALyGEB1pyGAQMLnpAwep3yVJKKYQIQbt7KsXA2EvFxdoFC2HBhSxD8puiKEZRzR6adtceUsKl2MLB77++gp0zoZqj1sq4tYyrYx1TR12l3StJ1JJSHrl9o5TyiBCi4f1OLKXMFUKMBX5DuwV2QUEF2VEF++cAA4DRQohc4DowSJZ24oaBzM3MqWFbQ7UkFKWiEAJq+miPNuPgxlU4/Xth11TMT9pxtQMKy4W4B4O5ISXnlLJyr+VLT0gp7xgfKNh3UkrpadTISvAwt8AOXTcUCzMLFnRdUMZRKYpSpqSE5GOFCePMbpB52uxvx7rgVA8c3bSxDCe34s+tVaXauzHGLbB7hRAvSim/ue2NXgD2lfaNKoJatrU4lnLHzVOKolQ0QkBtf+3R7jW4ng5xW7U1M66ehStntTuoMs5ryaOoao5astAnkSIPRzftoe6sMti9ksQEYJUQYgiFSaElYAX0NXZgxlDbrjZbErcgpUStm6QojxCbx8Cvr/YoKi8Xrl3QksaVMwUJJKnw+dl9cP32JV2FVkpdn0jqFbRG3MGxIJnY1Sj1uuGVVYlJQkqZjDaHoSPgX7D5FynllnKJzAhq2dYiOz+btJtpOFs7mzocRVEelrlFYSvhrtOw0OZt6JNHUsG/z2iJ5GKMNgkwJ+u28xZ0a91KGrcnkSrUrWXIjOutwNZyiMXoattpcyWSM5NVklCUqsLKtvA23LuREq6nFSaOK0lwNamwRWJot9Zj9bU5H7X8K1WV3Cp1m0AtW22uxIXMC/i4+Jg4GkVRKgQhwNZZe9QJuvsxhnZrbX4fXJuA/wDw71dyYnqEVKkkoW9JZKnbYBVFKQVDurUyL0P0Gji6ErZNg21ToXYg+PfXxlIe0dpVVWpkxtnaWZtQl2n4hLq9F/ayPWk72XnZRoys6lq/fj3e3t54enoyffr0O/anpaXRt29fAgMDCQkJ4ejRoyaIUlEMYOcKwSNhxC/wWjR0naaNbWyaAjMDYV4Y7JoDGQ83ofd+vzNXrlzhqaeeIigoCD8/PxYuXFhsvxDCXAhxQAix1pD3K/WiQ6b2sKXCuy7vSvNazZnWbtp9jz16+SjPr3ue3Pxc7CztaO/WntAGobRza4etpe0Dx6Bo8vLyaNKkCRs3bsTd3Z3g4GCWLFmCr2/helZvvPEG9vb2TJkyhdjYWMaMGcPmzZtNGLWilFJavNa6OLoSko8AQlusyb8/+PQCOxeDT2XI78zUqVO5cuUKH330EZcuXcLb25sLFy5QrVq1fVLKlkKI19DuVHU0ZJmGKtWSAK2GkyEtiSs3r/D6ttepYVODLzp+QbeG3dh1fheTfp9E+x/bM27LOH469RNXbl4ph6grp6ioKDw9PWnUqBFWVlYMGjSINWvWFDsmOjqaTp205dSbNm1KfHw8ycmqu1B5hFRvqM31GP0HjImCJ97UBsLXToBPm8B3A+DgErhx/78lhvzOCCHIyMhASsm1a9dwdnbGwsLi1j53oAcwz9Dwq9SYBGjVYI+m3LvLQkrJO3+8w8XrF4nsFklgjUBC64fyzj/e4cDFA2xO3MymhE1sPbMVC2FBcO1gOjfoTKf6nXC1cS2nK3n0nT17lnr1CmtAuru7s3v37mLHBAUFsXLlStq1a0dUVBQJCQkkJSVRq1at8g5XUR5eDW/o+BZ0iIALR+DoCq2FsXoUmFfT7o7y7w9Numl3Zd3GkN+ZsWPH0qtXL+rWrUtGRgY//vgjZoVzPr4A/gk4GBpylWxJJGcmU7SbLSkti58OndM/jzwWybakbUxqOYnAGoXLZliYaQkhIiSCDQM28EP3HxjqN5Sz187yn13/odP/OjFs3TAWH1vMuWvnUB5eREQE6enp6HQ6Zs2aRbNmzTA3Nzd1WIrycISAOoEQ9j5MOAwjN0HLcEjaA8tHwAwv2DpNq5pbSr/99hs6nY5z585x8OBBxo4dy9WrVwGcgItSylJVzKh6LQm72ndMqFux7yxfbD5Om8YuJGZF88X+LwhrEMbgpoNLPI+ZMCOgRgABNQKY0HwCJ9JPsDlhM5sSN/HJ3k/4ZO8n+Lr40rl+Z0IbhNLISa39ezs3NzfOnClclyopKQk3N7dixzg6OuoH3qSUeHh40KiR+l4qlYgQUC9Ye3T9EBL+hD3fwO/TYd9C6PgvaPYcmJkb9DuzcOFCIiIiEELg6emJh4cHsbGxAPZAr4K1e6wBRyHEd1LK5+4VXtVrSRSZK3FLB+8aSAnrYk4w6fdJ1LWvy/uPv29w6Q4hBE2qN2G0bjQreq3gl76/8FqL17AQFvz3wH/pvbo3vVf3ZtaBWcSkxPCo3SxgLMHBwZw4cYLTp0+TnZ3N0qVL6dWrV7Fj0tPTyc7W7iybN28e7du3x9Gxasx0VaogM3PwaAcDF8PIjfBYA/h5HMxpC5veJ9j1+n1/Z+rXr6+/uSM5OZm///771gers1JKdyllQ7SlG7bcL0FAFW1JgDbr2tdFuyMgwM0JZzsLvomZyjWRznfdv8PByuAuuzvUd6zPCP8RjPAfwYXMC2xJ3MLmxM3MOzKPuYfn4mbvRuf6nencoDNBNYIqdR2p3Lx8zl+5QXxKJvEpWSRc1r7ezM3Dt44jQ19/n9DOXRDkEx4ejp+fH3PmzAFg1KhRxMTEMGzYMIQQ+Pn5MX/+fBNfkaKUk3ohMHKDNvdiywfwx2dYyHy+/PJLunbtSl5e3l1/ZyZPnszw4cMJCAhASslHH32Eq+uDj5VWuVtgL2VdotOyTrzd6m2ebfqsfnuf79/nVO5y3mn1Ls80fbosQr1D6o1Utp3ZxqaETfx1/i9y83PxcfZhdNBoOtTr8Mgmi5y8fJLSrhOfkqlPAgkpmSSkZHEmLYucvMKfMWtLMxo422FpITh+4RrZefkAOFpb4O/mRICbE/5uTgS6O1Hf2faR/Z4oSpnLz3+oooPGKBVeKbnYuGBhVnxCXdT5KOJyV5JzRYeXTajR3tvZ2pl+Xv3o59WPjOwMNiZsZN6ReYzbOq7CJ4ubuXmcSb1OQkphErj1NSntOnn5hYnAzsqcBi52NK3jQFf/2jR0saWBix0NXeyo6VANMzPt+rJz8zmenMGRs1c4cvYKR89eYeHO+BITR4CbEw1cVOJQqigTVaWtci0JgG4ruqGrqWN6u+lcvn6ZAT8NwM7SgZi9I3i1oz8Tw5qUUbT3l5ufyy9xv/D14a85k3HGpMnienYeialZWougaDK4nMW5K9cp+qPiUM2Chq52NHCxpaGLHQ1d7ccvAPsAABm1SURBVPTJwNXe6oFjv1viiD2foU8cDtYW+NfVWhoqcSiK4R60JVElk8SwdVof9/wu83lx44scuXSEH3r8wD+XXCRfwpoxbcooWsPl5ueyNm4tcw/P1SeLV3Sv8IT7E2X6BzDzZi4JRVoC8ZczC5JCFheu3ih2bHVby4IWQEFLwLWwRVDd1rLc/jDfShxHz17h8D0SR0BB4ghUiUNR7qC6m0qhll0tjlw6wuxDs9lzYQ8ftPkAr+pedGgCX2w+Tsq1m7jYVyvXmCzMLOjj2YeejXqyNm4tXx/6mle3vPrQyeJKVg5fbD7O0bNXiE/J4lLGzWL7Xe2taOBix+OeLsVbBM52ONlaltXlPRQrCzP8C7qcBhVsK5o4brU4FhXpqro9cQS4OdHA2Vbf1aUoimGqZEvis72fERkdiZSSPp59+HebfwNwOCmdXl/u5PNngujbzL0swn1gOfk5WjfUoa9Jupb0QMliT3wq45cc4GLGTZo3qF5sbKCBiy0NXGxxsK4YiaAs3C1xxFzIIDu3IHFUs8DPzZFA98dU4lCqHNWSKIVadrXIl/l4VffirVZv6bf713XC1d6KbX9fMnmSsDSzpI9nH3o06sHaU1o31KtbXsXXxZdX/r+9O4+uujwTOP59skESCElMAuQCJnizkAXCkmo7gFSqMFZA1FaLjFWkDigW/nCm9HA4x9k8Yh1aOjjHggU7WpapiCCrGowM1Aqo7ASSAELCFswOiSQ37/zxu7ncQAIJyV1Cns85Ocn93fd375OXkCfvPuR5Rvcb3WKycDQYFm8rYFHOMfpHh7Fm5g8Y0j/Sy9+B9zXX4qhzuHVVFTlbHH89eV3iuDqrKlITh1JuumSSGBY3jJSoFH5z728IDQp1XQ8IEEYnxfLp0Qs4GgyBfvCLIjggmMlJk3norodcyWLWtlktJouzFTXMXrWXXSdKeTgrnn97OOO2ai20VXBgAOnxvUiP78Xj2dY198RhDZBX8qfPv2kxcWTaepFwR7gmDtUldcnuphtZt7eY2av2svb5HzB0QJTH3udW1TXUsaFwA3/Y/weKq4ubJIuPDp/nV2v2c6W+gX9/OINHhvm2NdSZNJc4jpytbDFxZPWP1HUcqlPR2U0dpOzSFYb/+8e8eF+SV6fCttW1ySIycCBnTo4mtVc2i382nMSYcF+H2OnVORrIP1/tnFVVfl3iiO3ZjeyEKEbcGU12QjSD+vYkKLDL7XSjOglNEh1o8n/v9NlU2LY6fLaMf3z/D5SGbCYgpJS06HReGPo8o2yj9K9cD2hMHF+fLmPPyTJ2nyylqKwGgLCQQIYNiGJEQhTZCdEMHRBJWEiX7NFVfkiTRAf6fU4+v/3kGHvm/cjrU2FbyxjDyl2n+dcNh+jRLYgFj6ZTGfSFq2WRfkc6z2fdWrLYsmULs2fPxuFwMH36dObOndvk+bKyMqZNm0ZhYSHdu3dn2bJlZGRkdOS316mcrahhz8ky9pwsZffJMo6cq8QYCAwQMuIjGJEQTXZCFMPvjCa2p3/+PKnbnyaJDuRPU2GbU3G5jrnv72fzwXOMSorhP386hLie3QGrG+rDwg9Zsn8JxdXFZNyRwcysma1OFnqkaPtV1tbx1TdXWxp7T5fznbOLKjEmnBF3Wi2NEQlRJMaEa4tPeYVOge1A/jQV9lruax9+/fep/GLUwCazboIDgnkk6REm3DXBlSxeyHmh1cnC/XhEwHU8onuSOHz4sKt14X6kqJ4WZ4noHsyYlDjGpMQB1vqNg2cqXC2NT46c5y9fFgHWYsYRd0a7uqjS4iMI1nEN5Uc8miREZDywCAgE3jLGvNpCuWzgc+AJY8x7noypNfxxKmxb1z64ksXACawvXM/SA0t5IecFMmMymTlkJiNtI5tNFnqkaMcLCQpg2IAohg2I4rnR0NBgOH6xmt3Olsaek2VsOWRtOBkaHMjQAZGuLqqhA6Lo0c13f8vdrOuxoqKCqVOncurUKerr63nppZd45plnXM87HA5GjBiBzWZjw4YN3g5fdQCP/fSJSCDwBnA/UATsFpH1xpjDzZRbAHzkqVhuxb0psbz/dTH7i8p9PhW2PWsfggODeTT5USbeNZH1hetZsn8Jz+c8f9NkcSNz585l9uzZZGVlkZmZqUeKtlFAgGCP64k9ric/+94AAM5X1rq6p3afLGXxtnwaDAQIpPZxTr3tZ63ZSO3Tk+7Bnq9vh8PBCy+80KTrceLEiU1alW+88QZpaWl8+OGHlJSUkJKSwpNPPklISAgAixYtYtCgQY3HZ6pOyJN/onwPKDDGHAcQkVXAJODwNeVeBNYA2R6Mpc1GJ8USIJB7tMSnSeKjQ+f4Z+fah4U/HXLLax9amyz0SFHf6B3RnR8P7suPB/cFoKq2jq9PlbvGND46fI7Ve6x/l8AAISmuB5m2q3tTpfWN6PDE0ZquRxGhqqoKYwzV1dVER0cTFGT9WikqKmLjxo3MmzePhQsXdmhsyns8mSRswGm3x0XA3e4FRMQGTAZ+iJ8liajwELL6R5J7rMQn6yVq6xy8sukI//P5N2TYIvivnw3rkLUP7sliXeE6lu5fyvM5zzM4ZjAzhszg7hF3u45HtNlsrFq1ihUrVjR5jfLycsLCwggJCdEjRT2kZ/dgRifHMjo5FrCScXF5jdu+VJVsy7vgGttoTBzp8b3ItEWQ2a8XaX17ERpy64mjNV2Ps2bNYuLEicTHx1NVVcXq1asJcJ57MGfOHF577TWqqqpuOQble74euP4d8CtjTMONujxE5DngObDOb/WWMSlx/PYT7+8Km3++ihdXfk3euSp+MSqRfxqXSkhQxw5mBgcG81jyY0y6a1KTZBEWFMaApwYw4t4RBBHE5CmT6TOwjx4p6mMiQr+oMPpFhTE+w2ptGGM4W1Hr2szwYHEFnx0rYc1XVuIIELDH9SAj3rm9SD+rxRHegWMcW7duJSsri23btlFYWMj999/PqFGj2L59O3FxcQwfPpzc3NwOez/lfZ5MEsVAf7fH/ZzX3I0AVjkTRAzwoIjUG2M+cC9kjFkCLAFrCqzHIr7GmJRYFn58jO353pnldO3ah7efyXbNkPEU92Sx9ZutHLx4kPyYfBpSGyj7royd7GTM/44hOiqapMgkFuxagP0OO+/tfA97pJ2w4DCPxqdaJiLER4YSHxnKuHTr7HZjDOcrv3Md2nSouIIdBRd5/+ti5z3WNNyBMT24Kzbc+jq2BwNjw7kjvOlhUa3pely+fDlz585FRLDb7SQmJpKXl8fOnTtZv349mzZtora2lsrKSqZOncq7777rhZpRHclj6yREJAg4BozFSg67gSnGmEMtlH8b2HCz2U3eWCfRqKHB8L1XPiEwQPj+wDtIi48grW8vBvXt2eEti4rLdfx67X42Hbh+7YMvGGP4tvZb8svyKSgvcH0uKC+gpr7GVc7Ww0ZSZBJJUUnYI+0kRSWREJFAcGDX3VTQH12orHUljryzVRy/WM3Ji5dd52+AdQbHwFgrecz6oZ0BUd1JTk4mJycHm81GdnY2K1asID093XXPzJkz6d27Ny+//DLnz59n2LBh7Nu3j5iYGFeZ3NxcXn/9dZ3d5GN+t07CGFMvIrOArVhTYJcZYw6JyAzn82966r07SkCA8C8TM1jzVRF/O17KB3vPuJ7rHdGNtL4RTRLHre4UuudkKbNX7eV8ZW2zax98QUSICY0hJjSG78d/33W9wTRQXF18XfLYUbyDelMPQJAEkdAr4brkEd8jngDRNQC+EBfRnbER3Rk76Oo0ZUeD4Ux5DYUl1RwvucSJi5c4frGarQfPceRsFR/O+jsWL17MuHHjcDgcTJs2jfT09CZdj/Pnz+fpp58mMzMTYwwLFixokiBU56crrtug9NIVjpyt5PCZSg6ftTZ7y79QjaPBqsOwkEBS+/RkkCt5RJDaJ6LFwUNHg+GNTwv43SfW2offPzG00577UOeo40TlieuSR3H11R7GsKAw7JF27FF2kiKTXJ/vCL3Dh5Gra205eJYl24/z5j8M92lrVnUs3ZbDR2rrHBRcqOawe/I4U0nVd9Zf1Y19wI2tjkF9I0jvG4HDGOas2ssXt/m5D9VXqimsKGySPPLL8in7rsxVJrp7dJOkYY+yY4+0Ex6sO9n6gjEGY/B5a1Z1LE0SfsQYQ1FZjStxHDlrJY/G3ULBmnnSPTiwy577cLHmYpMWR+Pn5sY73JNHYkSijncodQs0SXQCFTV1HHF2U50pr2HK3XfquQ9uGsc7CsoKyC/Pd30+WXGy2fEO9+Rh62HT8Q6lbkCThLptNY53XJs83Mc7QoNCXQPk7p9jQnUQVSnQJKG6oEt1l6xpuc6k0dJ4x7XJIzkqucnZ5kp1BX43BVYpTwsPDmdI7BCGxA5pcr1xvMO95fF+/vuu8Y6IkAimDJrCk6lPEtm9c84mU8pbtCWhuoQG08CZ6jMcKzvG+sL15JzKITQolJ8k/4Sn0p6id7huc65ub9rdpFQbFJQVsOzgMjad2ESABDDJPolp6dPoH9H/5jcr1QlpklDqFhRVFfH2obdZm7+WelPP+ITxPJv5LMlR3t/5VylP0iShVDuUXC7hncPvsProai7XX2ZM/zFMz5x+3XiHUp2VJgmlOkDFdxWsyFvBn4/8mYrvKri7z91MHzydu/vc3eYT/JTyJ5oklOpAl+su85djf+FPh/5ESU0JmTGZTM+czpj+Y3TRnuqUNEko5QFXHFdYV7iOZQeWUVRdhD3SzrOZzzI+YTxBATqDXHUemiSU8qD6hnq2ntzKWwfeoqC8AFsPG9MypjHJPolugd47tVCpW6VJQikvaDANfHb6M5YeWMqBiweIDY3l5+k/5yfJP9FT+pRf0yShlBcZY9h1bhdLDyzli7NfEBESwdRBU5kyaAq9uvXydXhKXUeThFI+sr9kP28deItPT39KaFAoj6c8zlNpTxEbFuvr0JRy0SShlI/ll+Xzx4N/ZPOJzQRKIJPtk3k642n699RV3Mr3NEko5SdOV51m+cHlfFDwAQ2mgfGJ43k241mSopJ8HZrqwjRJKOVnLly+4FrFXVNfw33972N65nQyYzN9HZrqgjRJKOWnymvLWZm3knePvEvllUru6XsPv8j8Bdl9snUVt/IaTRJK+blLdZd479h7vH3obS7WXGRw7GCmZ0zn3v736ipu5XG3miT0J1MpLwkPDufn6T9ny6NbmH/PfL6t+ZZffvpLJn4wkXk75vHO4XfYfW43lVcqvR7bli1bSElJwW638+qrr173fEVFBRMmTGDIkCGkp6ezfPnyJs87HA6GDh3KQw895K2QlZdoS0IpH6lvqGfzic1sPLGRo6VHuVhz0fWcrYeN1OhUUqJTGBQ9iNToVHqH9fZI95TD4SA5OZmPP/6Yfv36kZ2dzcqVK0lLS3OVeeWVV6ioqGDBggWUlJSQkpLCuXPnCAkJAWDhwoXs2bOHyspKNmzY0OExqvbT40uV6mSCAoKYcNcEJtw1AbCOXc0rzXN9HC09yrZT2zBYf8hFdoskJTqF1KirySOhV0K795DatWsXdrudgQMHAvDEE0+wbt26JklCRKiqqsIYQ3V1NdHR0QQFWe9bVFTExo0bmTdvHgsXLmxXLMr/aJJQyk/EhMYw0jaSkbaRrmuX6y5zrOxYk+SxMm8lVxquANAtsBtJkUlW8ohOJTU6leSo5DZtEVJcXEz//lfXcvTr148vvviiSZlZs2YxceJE4uPjqaqqYvXq1QQEWL3Vc+bM4bXXXqOqqqo9377yU5oklPJjYcFhZMVlkRWX5bpW31DPiYoTrtZGXmkeH3/zMWvy1wAgCBkxGdw34D7GDhhLYq/EdsexdetWsrKy2LZtG4WFhdx///2MGjWK7du3ExcXx/Dhw8nNzW33+yj/o0lCqU4mKCCIpKgkkqKSXF1VxhjOXTpHXmkeh0sPs6NoB4u+WsSirxYxsNdAxg4Yy9g7x5IWnXbduIbNZuP06dOux0VFRdhstiZlli9fzty5cxER7HY7iYmJ5OXlsXPnTtavX8+mTZuora2lsrKSqVOn8u6773q+IpRX6MC1Urepc5fOse3UNnJO5fDl+S9xGAd9w/u6WhjhweFcuHyBkX1HkpycTE5ODjabjezsbFasWEF6errrtWbOnEnv3r15+eWXOX/+PMOGDWPfvn3ExMS4yuTm5vL666/rwLWf8suBaxEZDywCAoG3jDGvXvP8JODfgAagHphjjNnhyZiU6ir6hPdhyqApTBk0hfLacnKLcsk5lcN7x97jz0f+DEB092g2P7KZxYsXM27cOBwOB9OmTSM9PZ0333wTgBkzZjB//nyefvppMjMzMcawYMGCJglC3b481pIQkUDgGHA/UATsBn5mjDnsVqYHcMkYY0RkMPC/xpjUG72utiSUap/LdZfZUbyDry58xYOJDzI4drCvQ1Je4I8tie8BBcaY4wAisgqYBLiShDGm2q18ONC5+r6U6oTCgsN4IOEBHkh4wNehqE7AkyuubcBpt8dFzmtNiMhkEckDNgLTmnshEXlORPaIyJ6SkhKPBKuUUup6Pt+Wwxiz1tnF9DDW+ERzZZYYY0YYY0bExupBLkop5S2eTBLFgPtpK/2c15pljNkODBQRHQ1TSik/4ckksRtIEpFEEQkBngDWuxcQEbs4J22LyDCgG/CtB2NSSinVBh4buDbG1IvILGAr1hTYZcaYQyIyw/n8m8CjwFMiUgfUAI+bzrZwQymlbmO6mE4ppboAPU9CKaVUh9MkoZRSqkWdrrtJREqAb27x9hjg4k1L+QeN1TM0Vs/QWD2jI2O90xjT5jUEnS5JtIeI7LmVPjlf0Fg9Q2P1DI3VM/whVu1uUkop1SJNEkoppVrU1ZLEEl8H0AYaq2dorJ6hsXqGz2PtUmMSSiml2qartSSUUkq1wW2RJERkvIgcFZECEZnbzPMiIr93Pr/fuU9Uq+71s1hPisgBEdkrIh5fdt6KWFNF5HMR+U5EXmrLvX4Wq7/V65POf/sDIvJXERnS2nv9LFZ/q9dJzlj3Oo8eGNnae/0sVq/WK8aYTv2BtS9UITAQCAH2AWnXlHkQ2AwIcA/wRWvv9ZdYnc+dBGL8qF7jgGzgP4CX2nKvv8Tqp/X6AyDK+fXf+/nPa7Ox+mm99uBqF/tgIM+P67XZWL1dr8aY26Il4ToBzxhzBWg8Ac/dJOB/jOVvQKSI9G3lvf4Sq7fdNFZjzAVjzG6grq33+lGs3taaWP9qjClzPvwb1jb7rbrXj2L1ttbEWm2cv2VpehKmP9ZrS7F63e2QJFpzAl5LZVp1el4Hak+sYP2gfCIiX4rIcx6L8uZxePLeW9He9/Pnen0Wq2V5K/e2V3tiBT+sV2n+JEy/rNcWYgXv1qtHz7hWHW+kMaZYROKAj0Ukz1iHNan28ct6FZEfYv3iHXmzsr7WQqx+V6/GmLXAWhEZjXUS5o98Gc+N3CBWr9br7dCSaM0JeC2VadPpeR2gPbFijGn8fAFYi9Vs9ZT21I0/1muL/LFeRWQw8BYwyRjzbVvu7UDtidUv69UtNveTMP2yXhtdE6u36/W2GLgOAo4DiVwdBEq/psyPaToYvKu19/pRrOFAT7ev/wqM92WsbmVfpunAtd/V6w1i9bt6BQYABcAPbvX79INY/bFe7VwdDB6G9YtZ/LReW4rVq/VqjOn8ScJZWQ8Cx7BmDMxzXpsBzHB+LcAbzucPACNudK8/xoo1E2Kf8+OQn8TaB6s/tRIod34d4af12mysflqvbwFlwF7nxx4//nltNlY/rddfOWPZC3yO1W3jr/XabKy+qFddca2UUqpFt8OYhFJKKQ/RJKGUUqpFmiSUUkq1SJOEUkqpFmmSUEop1SJNEqrLEZE+IrJKRAqdWxtsEpFkX8ellD/SJKG6FBERrFWqucaYu4wxw4FfA7098F667Y3q9DRJqK7mh0CdMebNxgvGmH3ADhH5jYgcdO7V/ziAs8Xx48ayIvK2iDwmIoHO8rud+/7/o/P5MSLyfyKyHjjsvPaBs8VyyH1DNhF5VkSOicguEVkqIoud12NFZI3ztXeLyN95pWaUaob+paO6mgzgy2auPwJkAUOAGGC3iGwHVgM/BTaKSAgwFpiJtZldhTEmW0S6ATtF5CPnaw0DMowxJ5yPpxljSkUk1Pm6a4BuwHxn2SpgG9YqWoBFwG+NMTtEZACwFRjUcVWgVOtpklDKMhJYaYxxAOdF5DOsQ4o2A4uciWA8sN0YUyMiDwCDReQx5/29gCTgCtZ+WyfcXvuXIjLZ+XV/Z7k+wGfGmFIAEfkL0Dgu8iMgzeoZAyBCRHoYY6o7/ttW6sY0Saiu5hDw2E1LORljakUkFxgHPI51QAxYe2y9aIzZ6l5eRMYAl655/CPg+8aYy87X6n6Ttw0A7jHG1LY2TqU8RcckVFezDeh2zdjAYKxN/x53jjXEAqOBXc4iq4FngFHAFue1rcBMEQl2vkayiIQ38369gDJngkjF2tkXYDdwr4hEOQe4H3W75yPgRbf4str1HSvVDtqSUF2KMcY4u35+JyK/Amqxzgyeg3Wu8D6sk7/+2RhzznnbR8A7wDpjHTcJ1u6nCcBXzhlTJcDDzbzlFmCGiBwBjmId8YmxDo15BSsRlQJ5QIXznl8Cb4jIfqz/o9uxdghVyut0F1ilfKRxnMHZklgLLDPWaWRK+Q3tblLKd14Wkb3AQeAE8IGP41HqOtqSUEop1SJtSSillGqRJgmllFIt0iShlFKqRZoklFJKtUiThFJKqRZpklBKKdWi/weDgdBcuVof6gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f475fbb5490>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1,1,squeeze=False)\n",
    "for ky, res in final.items():\n",
    "    plot_curve(axes[0,0], res, {'x': 'coverage', 'y': 'clus_purity'}, label=ky, annot_freq=6)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 78),\n",
       "              ('n_clus', 37),\n",
       "              ('mean_len', 18.102564102564102),\n",
       "              ('mean_n_type', 1.1923076923076923),\n",
       "              ('ned', 0.67907801418439706),\n",
       "              ('coverage', 0.015837129775848946),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.54947526236881561),\n",
       "              ('inv_purity', 0.63493253373313341),\n",
       "              ('garbage_ratio', 0.10869565217391304),\n",
       "              ('avg_purity', 0.6538461538461539),\n",
       "              ('clus_purity', 0.62820512820512819),\n",
       "              ('clus_purity_inv', 0.74358974358974361),\n",
       "              ('grouping_P', 0.19230769230769232),\n",
       "              ('grouping_R', 0.35714285714285715),\n",
       "              ('grouping_F', 0.25000000000000006),\n",
       "              ('type_P', 0.7777777777777778),\n",
       "              ('type_R', 0.020289855072463767),\n",
       "              ('type_F', 0.03954802259887005),\n",
       "              ('dtw', 0.89000000000000001),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 156),\n",
       "              ('n_clus', 76),\n",
       "              ('mean_len', 18.012820512820515),\n",
       "              ('mean_n_type', 1.1602564102564104),\n",
       "              ('ned', 0.80952380952380953),\n",
       "              ('coverage', 0.031197100691880347),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.53428786737000755),\n",
       "              ('inv_purity', 0.60813865862848526),\n",
       "              ('garbage_ratio', 0.14694800301431801),\n",
       "              ('avg_purity', 0.6089743589743589),\n",
       "              ('clus_purity', 0.5641025641025641),\n",
       "              ('clus_purity_inv', 0.75),\n",
       "              ('grouping_P', 0.14102564102564102),\n",
       "              ('grouping_R', 0.3384615384615385),\n",
       "              ('grouping_F', 0.1990950226244344),\n",
       "              ('type_P', 0.75),\n",
       "              ('type_R', 0.02608695652173913),\n",
       "              ('type_F', 0.050420168067226885),\n",
       "              ('dtw', 0.90000000000000002),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 280),\n",
       "              ('n_clus', 121),\n",
       "              ('mean_len', 17.875),\n",
       "              ('mean_n_type', 1.1785714285714286),\n",
       "              ('ned', 0.82033333333333347),\n",
       "              ('coverage', 0.055736699196782584),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.52740740740740744),\n",
       "              ('inv_purity', 0.47428571428571431),\n",
       "              ('garbage_ratio', 0.13544973544973546),\n",
       "              ('avg_purity', 0.575),\n",
       "              ('clus_purity', 0.53928571428571426),\n",
       "              ('clus_purity_inv', 0.59999999999999998),\n",
       "              ('grouping_P', 0.15357142857142858),\n",
       "              ('grouping_R', 0.2529411764705882),\n",
       "              ('grouping_F', 0.1911111111111111),\n",
       "              ('type_P', 0.7894736842105263),\n",
       "              ('type_R', 0.043478260869565216),\n",
       "              ('type_F', 0.08241758241758242),\n",
       "              ('dtw', 0.91000000000000003),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 1205),\n",
       "              ('n_clus', 155),\n",
       "              ('mean_len', 19.513692946058093),\n",
       "              ('mean_n_type', 1.1692946058091287),\n",
       "              ('ned', 0.97696289506889178),\n",
       "              ('coverage', 0.24289658149759716),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.20287776233806984),\n",
       "              ('inv_purity', 0.61593975525572642),\n",
       "              ('garbage_ratio', 0.17732753597202922),\n",
       "              ('avg_purity', 0.23070539419087138),\n",
       "              ('clus_purity', 0.2033195020746888),\n",
       "              ('clus_purity_inv', 0.73195020746887962),\n",
       "              ('grouping_P', 0.4854771784232365),\n",
       "              ('grouping_R', 0.6565656565656566),\n",
       "              ('grouping_F', 0.558206106870229),\n",
       "              ('type_P', 0.8125),\n",
       "              ('type_R', 0.03768115942028986),\n",
       "              ('type_F', 0.07202216066481995),\n",
       "              ('dtw', 0.92000000000000004),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 653),\n",
       "              ('n_clus', 129),\n",
       "              ('mean_len', 19.689127105666156),\n",
       "              ('mean_n_type', 1.1761102603369067),\n",
       "              ('ned', 0.93711281216858044),\n",
       "              ('coverage', 0.12799218368343918),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.32718780727630287),\n",
       "              ('inv_purity', 0.56489675516224191),\n",
       "              ('garbage_ratio', 0.17969518190757128),\n",
       "              ('avg_purity', 0.37059724349157736),\n",
       "              ('clus_purity', 0.33078101071975496),\n",
       "              ('clus_purity_inv', 0.68606431852986216),\n",
       "              ('grouping_P', 0.4226646248085758),\n",
       "              ('grouping_R', 0.6146993318485523),\n",
       "              ('grouping_F', 0.5009074410163339),\n",
       "              ('type_P', 0.7647058823529411),\n",
       "              ('type_R', 0.03768115942028986),\n",
       "              ('type_F', 0.07182320441988951),\n",
       "              ('dtw', 0.93000000000000005),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 303),\n",
       "              ('n_clus', 74),\n",
       "              ('mean_len', 19.079207920792079),\n",
       "              ('mean_n_type', 1.099009900990099),\n",
       "              ('ned', 0.89965149651496668),\n",
       "              ('coverage', 0.056225218981833879),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.37897042716319823),\n",
       "              ('inv_purity', 0.57776560788608977),\n",
       "              ('garbage_ratio', 0.16611902154070829),\n",
       "              ('avg_purity', 0.40264026402640263),\n",
       "              ('clus_purity', 0.37623762376237624),\n",
       "              ('clus_purity_inv', 0.65346534653465349),\n",
       "              ('grouping_P', 0.38283828382838286),\n",
       "              ('grouping_R', 0.5603864734299517),\n",
       "              ('grouping_F', 0.4549019607843138),\n",
       "              ('type_P', 0.8333333333333334),\n",
       "              ('type_R', 0.028985507246376812),\n",
       "              ('type_F', 0.05602240896358543),\n",
       "              ('dtw', 0.94000000000000006),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 108),\n",
       "              ('n_clus', 33),\n",
       "              ('mean_len', 19.851851851851851),\n",
       "              ('mean_n_type', 1.0740740740740742),\n",
       "              ('ned', 0.91143654114365413),\n",
       "              ('coverage', 0.019552152327285535),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.41895874263261296),\n",
       "              ('inv_purity', 0.46512770137524556),\n",
       "              ('garbage_ratio', 0.1556974459724951),\n",
       "              ('avg_purity', 0.42592592592592593),\n",
       "              ('clus_purity', 0.42592592592592593),\n",
       "              ('clus_purity_inv', 0.5),\n",
       "              ('grouping_P', 0.3333333333333333),\n",
       "              ('grouping_R', 0.47368421052631576),\n",
       "              ('grouping_F', 0.391304347826087),\n",
       "              ('type_P', 1.0),\n",
       "              ('type_R', 0.011594202898550725),\n",
       "              ('type_F', 0.022922636103151862),\n",
       "              ('dtw', 0.95000000000000007),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 16),\n",
       "              ('n_clus', 7),\n",
       "              ('mean_len', 21.1875),\n",
       "              ('mean_n_type', 0.9375),\n",
       "              ('ned', 0.79166666666666663),\n",
       "              ('coverage', 0.0030220061121777756),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.49535603715170279),\n",
       "              ('inv_purity', 0.46130030959752322),\n",
       "              ('garbage_ratio', 0.37461300309597523),\n",
       "              ('avg_purity', 0.5625),\n",
       "              ('clus_purity', 0.5625),\n",
       "              ('clus_purity_inv', 0.5),\n",
       "              ('grouping_P', 0.25),\n",
       "              ('grouping_R', 0.4444444444444444),\n",
       "              ('grouping_F', 0.32),\n",
       "              ('type_P', 1.0),\n",
       "              ('type_R', 0.005797101449275362),\n",
       "              ('type_F', 0.011527377521613832),\n",
       "              ('dtw', 0.96000000000000008),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')]),\n",
       " OrderedDict([('n_match', 113137),\n",
       "              ('n_node', 4),\n",
       "              ('n_clus', 2),\n",
       "              ('mean_len', 22.0),\n",
       "              ('mean_n_type', 0.75),\n",
       "              ('ned', 1.0),\n",
       "              ('coverage', 0.00084070846729757673),\n",
       "              ('matching_P', 0),\n",
       "              ('matching_R', 0),\n",
       "              ('matching_F', 0),\n",
       "              ('purity', 0.39285714285714285),\n",
       "              ('inv_purity', 0.42857142857142855),\n",
       "              ('garbage_ratio', 0.5714285714285714),\n",
       "              ('avg_purity', 0.5),\n",
       "              ('clus_purity', 0.5),\n",
       "              ('clus_purity_inv', 0.5),\n",
       "              ('grouping_P', 0.0),\n",
       "              ('grouping_R', 0),\n",
       "              ('grouping_F', 0.0),\n",
       "              ('type_P', 0),\n",
       "              ('type_R', 0.0),\n",
       "              ('type_F', 0.0),\n",
       "              ('dtw', 0.97000000000000008),\n",
       "              ('exp',\n",
       "               'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032')])]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['AE (Signer04 Signer05 Signer09)', 'AE (Signer04)', 'AE (Signer04)2']"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resname = 'train_set'\n",
    "train = load_obj(name='grid_post_disc_' + resname, path='/home/korhan/Dropbox/tez/files/ZR_dtw_curves')\n",
    "train.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f47ba2052d0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f47b665cc10>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "tmp = train\n",
    "ax_dict = {'x': 'coverage', 'y': 'clus_purity'}\n",
    "fig, axes = plt.subplots(1,1,squeeze=False, figsize=(5,5))\n",
    "\n",
    "\n",
    "ky = 'AE (Signer04)'\n",
    "res = tmp[ky]\n",
    "lb = 'AutoEncoder (4)'\n",
    "plot_curve(axes[0,0], slct_res(res[:],[]), ax_dict, '--x', label=lb, annot_freq=1)\n",
    "\n",
    "\n",
    "ky = 'AE (Signer04)2'\n",
    "res = tmp[ky]\n",
    "lb = 'AutoEncoder (4)2'\n",
    "plot_curve(axes[0,0], slct_res(res[:],[]), ax_dict, '--x', label=lb, annot_freq=1)\n",
    "\n",
    "\"\"\"ky = 'AE (Signer05)'\n",
    "res = tmp[ky]\n",
    "lb = 'AutoEncoder (5)'\n",
    "plot_curve(axes[0,0], slct_res(res[:],[]), ax_dict, '--x', label=lb, annot_freq=1)\n",
    "\"\"\"\n",
    "\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['AE (Signer08)',\n",
       " 'AE (Signer01,Signer08)',\n",
       " 'DH (Signer01,Signer08)',\n",
       " 'AE (Signer01 Signer08)',\n",
       " 'AE (Signer01,Signer08)3',\n",
       " 'AE (Signer01,Signer08)2',\n",
       " 'DH (Signer08)',\n",
       " 'AE (Signer01)',\n",
       " 'OP (Signer08)',\n",
       " 'OP (Signer01,Signer08)',\n",
       " 'DH (Signer01)']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resname = 'test_set'\n",
    "test = load_obj(name='grid_post_disc_' + resname, path='/home/korhan/Dropbox/tez/files/ZR_dtw_curves')\n",
    "test.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f3e98655850>"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f3e987dc0d0>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "tmp = test\n",
    "ax_dict = {'x': 'coverage', 'y': 'clus_purity'}\n",
    "fig, axes = plt.subplots(1,1,squeeze=False, figsize=(5,5))\n",
    "\n",
    "ky = 'DH (Signer01,Signer08)'\n",
    "res = tmp[ky]\n",
    "lb = 'DeepHand (test)'\n",
    "plot_curve(axes[0,0], res[:-2], ax_dict, '-x', label=lb, annot_freq=-1)\n",
    "\n",
    "ky = 'AE (Signer01,Signer08)'\n",
    "res = tmp[ky]\n",
    "lb = 'AutoEncoder (test)'\n",
    "plot_curve(axes[0,0], slct_res(res[2:-4],[]), ax_dict, '--x', label=lb, annot_freq=-1)\n",
    "\n",
    "ky = 'AE (Signer08)'\n",
    "res = tmp[ky]\n",
    "lb = 'AutoEncoder (S08)'\n",
    "plot_curve(axes[0,0], slct_res(res[2:],[]), ax_dict, '--x', label=lb, annot_freq=-1)\n",
    "\n",
    "ky = 'DH (Signer08)'\n",
    "res = tmp[ky]\n",
    "lb = 'DeepHand (S08)'\n",
    "plot_curve(axes[0,0], slct_res(res[:],[]), ax_dict, '-x', label=lb, annot_freq=-1)\n",
    "\n",
    "ky = 'DH (Signer01)'\n",
    "res = tmp[ky]\n",
    "lb = 'DeepHand (S01)'\n",
    "plot_curve(axes[0,0], slct_res(res[:],[]), ax_dict, '-x', label=lb, annot_freq=-1)\n",
    "\n",
    "ky = 'OP (Signer01,Signer08)'\n",
    "res = tmp[ky]\n",
    "lb = 'OpenPose (test)'\n",
    "plot_curve(axes[0,0], res[6:], ax_dict, '-.x', label=lb, annot_freq=-1)\n",
    "\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['AE (Signer03,Signer07)', 'OP (Signer03,Signer07)', 'DH (Signer03,Signer07)']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "resname = 'dev_set'\n",
    "dev = load_obj(name='grid_post_disc_' + resname, path='/home/korhan/Dropbox/tez/files/ZR_dtw_curves')\n",
    "dev.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f1b3be81410>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1b3bc3ce10>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "tmp = dev\n",
    "ax_dict = {'x': 'coverage', 'y': 'ned'}\n",
    "fig, axes = plt.subplots(1,1,squeeze=False, figsize=(5,5))\n",
    "\n",
    "ky = 'DH (Signer03,Signer07)'\n",
    "res = tmp[ky]\n",
    "lb = 'DeepHand (dev)'\n",
    "plot_curve(axes[0,0], res[4:-3], ax_dict, '-x', label=lb, annot_freq=-1)\n",
    "\n",
    "\n",
    "ky = 'DH (Signer01,Signer08)'\n",
    "res = test[ky]\n",
    "lb = 'DeepHand (test)'\n",
    "plot_curve(axes[0,0], res[1:-2], ax_dict, '-x', label=lb, annot_freq=-1)\n",
    "\n",
    "\n",
    "\"\"\"ky = 'AE (Signer03,Signer07)'\n",
    "res = tmp[ky]\n",
    "lb = 'AutoEncoder (dev)'\n",
    "plot_curve(axes[0,0], res[4:-2], ax_dict, '--x', label=lb, annot_freq=-1)\n",
    "\"\"\"\n",
    "\"\"\"ky = 'OP (Signer03,Signer07)'\n",
    "res = tmp[ky]\n",
    "lb = 'OpenPose (dev)'\n",
    "plot_curve(axes[0,0], res[6:], ax_dict, '-.x', label=lb, annot_freq=-1)\n",
    "\"\"\"\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.900323\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/korhan/anaconda3/envs/zerospeech/lib/python2.7/site-packages/ipykernel_launcher.py:3: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "cost_thr = 0.20\n",
    "\n",
    "cost_val = matches_df.score.sort_values(ascending=False).values[round(len(matches_df) * cost_thr)]\n",
    "print(cost_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./post_disc zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_35_02_7_05_06_9_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_35_02_7_05_06_9_06_032.lsh64 0.909653\n",
      "Read 106968 nodes/tokens\n",
      "Read 1703 clusters\n",
      "38.3903647006% OVERLAP\n"
     ]
    }
   ],
   "source": [
    "#cost_val = 0.90\n",
    "results_path = post_disc(exp_name, cost_val)\n",
    "\n",
    "\" get nodes \"\n",
    "nodes_df = get_nodes_df(exp_name)\n",
    "\n",
    "\" get clusters \"\n",
    "clusters_list = get_clusters_list(exp_name)\n",
    "\n",
    "scores = evaluate(None, signer_id, matches_df, nodes_df, clusters_list,\n",
    "                  group=3, interp=False, boundary_th=0.5, fast_compute=False)\n",
    "\n",
    "scores['exp'] = exp_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('n_match', 178351),\n",
       "             ('n_node', 3742),\n",
       "             ('n_clus', 1703),\n",
       "             ('mean_len', 23.741314804917156),\n",
       "             ('mean_n_type', 1.7014965259219668),\n",
       "             ('ned', 0.83056896978702366),\n",
       "             ('coverage', 0.25875534208237205),\n",
       "             ('matching_P', 0.14544741742624717),\n",
       "             ('matching_R', 0.03261681832168447),\n",
       "             ('matching_F', 0.053284501175921435),\n",
       "             ('purity', 0.41049143340619049),\n",
       "             ('inv_purity', 0.087863404545347717),\n",
       "             ('garbage_ratio', 0.14634891536816377),\n",
       "             ('avg_purity', 0.5611972207375735),\n",
       "             ('clus_purity', 0.50347407803313737),\n",
       "             ('clus_purity_inv', 0.43639764831640832),\n",
       "             ('grouping_P', 0.07856761090326028),\n",
       "             ('grouping_R', 0.13486238532110092),\n",
       "             ('grouping_F', 0.09929078014184396),\n",
       "             ('type_P', 0.582089552238806),\n",
       "             ('type_R', 0.07317073170731707),\n",
       "             ('type_F', 0.13),\n",
       "             ('exp',\n",
       "              'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_35_02_7_05_06_9_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_35_02_7_05_06_9_06_032')])"
      ]
     },
     "execution_count": 177,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>exp</th>\n",
       "      <th>n_clus</th>\n",
       "      <th>n_node</th>\n",
       "      <th>coverage</th>\n",
       "      <th>matching_P</th>\n",
       "      <th>matching_R</th>\n",
       "      <th>matching_F</th>\n",
       "      <th>grouping_P</th>\n",
       "      <th>grouping_R</th>\n",
       "      <th>grouping_F</th>\n",
       "      <th>clus_purity</th>\n",
       "      <th>clus_purity_inv</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>zrroot0_Signer01_Signer08_c3_right_PCA50_10_3_...</td>\n",
       "      <td>432</td>\n",
       "      <td>1086</td>\n",
       "      <td>0.084741</td>\n",
       "      <td>0.319031</td>\n",
       "      <td>0.024023</td>\n",
       "      <td>0.044682</td>\n",
       "      <td>0.407919</td>\n",
       "      <td>0.593039</td>\n",
       "      <td>0.483361</td>\n",
       "      <td>0.617864</td>\n",
       "      <td>0.502762</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 exp  n_clus  n_node  \\\n",
       "0  zrroot0_Signer01_Signer08_c3_right_PCA50_10_3_...     432    1086   \n",
       "\n",
       "   coverage  matching_P  matching_R  matching_F  grouping_P  grouping_R  \\\n",
       "0  0.084741    0.319031    0.024023    0.044682    0.407919    0.593039   \n",
       "\n",
       "   grouping_F  clus_purity  clus_purity_inv  \n",
       "0    0.483361     0.617864         0.502762  "
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "select = ['exp','n_clus','n_node','coverage',\n",
    "          'matching_P','matching_R','matching_F',\n",
    "          'grouping_P','grouping_R','grouping_F',\n",
    "          'clus_purity', 'clus_purity_inv']\n",
    "\n",
    "pd.DataFrame([scores], columns=scores.keys())[select]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filename</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>score</th>\n",
       "      <th>unknown</th>\n",
       "      <th>file_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-2</td>\n",
       "      <td>72</td>\n",
       "      <td>98</td>\n",
       "      <td>0.921689</td>\n",
       "      <td>4.670115</td>\n",
       "      <td>1775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>03July_2010_Saturday_tagesschau_default-14</td>\n",
       "      <td>103</td>\n",
       "      <td>128</td>\n",
       "      <td>0.921689</td>\n",
       "      <td>4.670115</td>\n",
       "      <td>156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-2</td>\n",
       "      <td>74</td>\n",
       "      <td>98</td>\n",
       "      <td>0.915857</td>\n",
       "      <td>3.789952</td>\n",
       "      <td>1775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>25October_2010_Monday_tagesschau_default-4</td>\n",
       "      <td>158</td>\n",
       "      <td>184</td>\n",
       "      <td>0.915857</td>\n",
       "      <td>3.789952</td>\n",
       "      <td>1726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-2</td>\n",
       "      <td>78</td>\n",
       "      <td>98</td>\n",
       "      <td>0.924372</td>\n",
       "      <td>3.391010</td>\n",
       "      <td>1775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>10August_2009_Monday_heute_default-11</td>\n",
       "      <td>101</td>\n",
       "      <td>123</td>\n",
       "      <td>0.924372</td>\n",
       "      <td>3.391010</td>\n",
       "      <td>685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>29</td>\n",
       "      <td>66</td>\n",
       "      <td>0.920092</td>\n",
       "      <td>23.872665</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>05December_2009_Saturday_tagesschau_default-0</td>\n",
       "      <td>36</td>\n",
       "      <td>70</td>\n",
       "      <td>0.920092</td>\n",
       "      <td>23.872665</td>\n",
       "      <td>240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>18</td>\n",
       "      <td>50</td>\n",
       "      <td>0.903186</td>\n",
       "      <td>8.026211</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>09August_2009_Sunday_tagesschau_default-0</td>\n",
       "      <td>14</td>\n",
       "      <td>41</td>\n",
       "      <td>0.903186</td>\n",
       "      <td>8.026211</td>\n",
       "      <td>589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>17</td>\n",
       "      <td>54</td>\n",
       "      <td>0.914556</td>\n",
       "      <td>16.556871</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>25June_2010_Friday_tagesschau_default-0</td>\n",
       "      <td>10</td>\n",
       "      <td>45</td>\n",
       "      <td>0.914556</td>\n",
       "      <td>16.556871</td>\n",
       "      <td>1701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>16</td>\n",
       "      <td>46</td>\n",
       "      <td>0.913693</td>\n",
       "      <td>3.632980</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>17June_2010_Thursday_tagesschau_default-0</td>\n",
       "      <td>11</td>\n",
       "      <td>38</td>\n",
       "      <td>0.913693</td>\n",
       "      <td>3.632980</td>\n",
       "      <td>1209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>17</td>\n",
       "      <td>49</td>\n",
       "      <td>0.914637</td>\n",
       "      <td>5.679434</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>25October_2010_Monday_tagesschau_default-0</td>\n",
       "      <td>14</td>\n",
       "      <td>46</td>\n",
       "      <td>0.914637</td>\n",
       "      <td>5.679434</td>\n",
       "      <td>1720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>38</td>\n",
       "      <td>64</td>\n",
       "      <td>0.903906</td>\n",
       "      <td>6.125254</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>30January_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>38</td>\n",
       "      <td>68</td>\n",
       "      <td>0.903906</td>\n",
       "      <td>6.125254</td>\n",
       "      <td>2043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>26June_2010_Saturday_tagesschau_default-0</td>\n",
       "      <td>19</td>\n",
       "      <td>58</td>\n",
       "      <td>0.924526</td>\n",
       "      <td>33.653538</td>\n",
       "      <td>1769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>29January_2011_Saturday_tagesschau_default-0</td>\n",
       "      <td>24</td>\n",
       "      <td>60</td>\n",
       "      <td>0.924526</td>\n",
       "      <td>33.653538</td>\n",
       "      <td>1978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>23February_2010_Tuesday_heute_default-1</td>\n",
       "      <td>42</td>\n",
       "      <td>69</td>\n",
       "      <td>0.934854</td>\n",
       "      <td>39.421871</td>\n",
       "      <td>1504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>22November_2010_Monday_heute_default-1</td>\n",
       "      <td>61</td>\n",
       "      <td>95</td>\n",
       "      <td>0.934854</td>\n",
       "      <td>39.421871</td>\n",
       "      <td>1465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>23February_2010_Tuesday_heute_default-1</td>\n",
       "      <td>40</td>\n",
       "      <td>70</td>\n",
       "      <td>0.930481</td>\n",
       "      <td>33.443237</td>\n",
       "      <td>1504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>01June_2011_Wednesday_tagesschau_default-6</td>\n",
       "      <td>40</td>\n",
       "      <td>73</td>\n",
       "      <td>0.930481</td>\n",
       "      <td>33.443237</td>\n",
       "      <td>67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>23February_2010_Tuesday_heute_default-1</td>\n",
       "      <td>41</td>\n",
       "      <td>71</td>\n",
       "      <td>0.912161</td>\n",
       "      <td>37.533344</td>\n",
       "      <td>1504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>30May_2011_Monday_heute_default-13</td>\n",
       "      <td>67</td>\n",
       "      <td>106</td>\n",
       "      <td>0.912161</td>\n",
       "      <td>37.533344</td>\n",
       "      <td>2096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>23February_2010_Tuesday_heute_default-1</td>\n",
       "      <td>41</td>\n",
       "      <td>73</td>\n",
       "      <td>0.915484</td>\n",
       "      <td>29.277313</td>\n",
       "      <td>1504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>12April_2010_Monday_tagesschau_default-2</td>\n",
       "      <td>57</td>\n",
       "      <td>85</td>\n",
       "      <td>0.915484</td>\n",
       "      <td>29.277313</td>\n",
       "      <td>860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>23February_2010_Tuesday_heute_default-1</td>\n",
       "      <td>41</td>\n",
       "      <td>70</td>\n",
       "      <td>0.941494</td>\n",
       "      <td>21.379238</td>\n",
       "      <td>1504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>05December_2011_Monday_heute_default-4</td>\n",
       "      <td>64</td>\n",
       "      <td>82</td>\n",
       "      <td>0.941494</td>\n",
       "      <td>21.379238</td>\n",
       "      <td>248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1739</th>\n",
       "      <td>23February_2010_Tuesday_tagesschau_default-3</td>\n",
       "      <td>43</td>\n",
       "      <td>67</td>\n",
       "      <td>0.912950</td>\n",
       "      <td>14.888820</td>\n",
       "      <td>1519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1740</th>\n",
       "      <td>05October_2011_Wednesday_heute_default-17</td>\n",
       "      <td>31</td>\n",
       "      <td>65</td>\n",
       "      <td>0.912950</td>\n",
       "      <td>14.888820</td>\n",
       "      <td>379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1741</th>\n",
       "      <td>31May_2011_Tuesday_heute_default-24</td>\n",
       "      <td>4</td>\n",
       "      <td>32</td>\n",
       "      <td>0.916441</td>\n",
       "      <td>11.166643</td>\n",
       "      <td>2163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1742</th>\n",
       "      <td>24February_2011_Thursday_heute_default-7</td>\n",
       "      <td>13</td>\n",
       "      <td>42</td>\n",
       "      <td>0.916441</td>\n",
       "      <td>11.166643</td>\n",
       "      <td>1607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1743</th>\n",
       "      <td>31May_2011_Tuesday_heute_default-24</td>\n",
       "      <td>5</td>\n",
       "      <td>32</td>\n",
       "      <td>0.913738</td>\n",
       "      <td>23.135624</td>\n",
       "      <td>2163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1744</th>\n",
       "      <td>24February_2011_Thursday_heute_default-7</td>\n",
       "      <td>14</td>\n",
       "      <td>52</td>\n",
       "      <td>0.913738</td>\n",
       "      <td>23.135624</td>\n",
       "      <td>1607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1745</th>\n",
       "      <td>31May_2011_Tuesday_heute_default-24</td>\n",
       "      <td>9</td>\n",
       "      <td>33</td>\n",
       "      <td>0.908595</td>\n",
       "      <td>12.309090</td>\n",
       "      <td>2163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1746</th>\n",
       "      <td>24February_2011_Thursday_heute_default-7</td>\n",
       "      <td>26</td>\n",
       "      <td>59</td>\n",
       "      <td>0.908595</td>\n",
       "      <td>12.309090</td>\n",
       "      <td>1607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1747</th>\n",
       "      <td>31May_2011_Tuesday_heute_default-24</td>\n",
       "      <td>8</td>\n",
       "      <td>32</td>\n",
       "      <td>0.904950</td>\n",
       "      <td>4.587836</td>\n",
       "      <td>2163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1748</th>\n",
       "      <td>23February_2010_Tuesday_tagesschau_default-12</td>\n",
       "      <td>95</td>\n",
       "      <td>113</td>\n",
       "      <td>0.904950</td>\n",
       "      <td>4.587836</td>\n",
       "      <td>1515</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1749</th>\n",
       "      <td>24August_2010_Tuesday_tagesschau_default-3</td>\n",
       "      <td>73</td>\n",
       "      <td>95</td>\n",
       "      <td>0.905229</td>\n",
       "      <td>4.708822</td>\n",
       "      <td>1597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1750</th>\n",
       "      <td>01February_2011_Tuesday_tagesschau_default-3</td>\n",
       "      <td>152</td>\n",
       "      <td>183</td>\n",
       "      <td>0.905229</td>\n",
       "      <td>4.708822</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1751</th>\n",
       "      <td>24August_2010_Tuesday_tagesschau_default-3</td>\n",
       "      <td>69</td>\n",
       "      <td>93</td>\n",
       "      <td>0.932088</td>\n",
       "      <td>18.755789</td>\n",
       "      <td>1597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1752</th>\n",
       "      <td>23May_2011_Monday_tagesschau_default-2</td>\n",
       "      <td>124</td>\n",
       "      <td>152</td>\n",
       "      <td>0.932088</td>\n",
       "      <td>18.755789</td>\n",
       "      <td>1553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1753</th>\n",
       "      <td>24August_2010_Tuesday_tagesschau_default-3</td>\n",
       "      <td>63</td>\n",
       "      <td>94</td>\n",
       "      <td>0.907828</td>\n",
       "      <td>11.890221</td>\n",
       "      <td>1597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1754</th>\n",
       "      <td>03February_2011_Thursday_tagesschau_default-11</td>\n",
       "      <td>62</td>\n",
       "      <td>93</td>\n",
       "      <td>0.907828</td>\n",
       "      <td>11.890221</td>\n",
       "      <td>148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1755</th>\n",
       "      <td>01February_2011_Tuesday_tagesschau_default-3</td>\n",
       "      <td>152</td>\n",
       "      <td>184</td>\n",
       "      <td>0.900610</td>\n",
       "      <td>26.614849</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1756</th>\n",
       "      <td>23May_2011_Monday_tagesschau_default-2</td>\n",
       "      <td>126</td>\n",
       "      <td>152</td>\n",
       "      <td>0.900610</td>\n",
       "      <td>26.614849</td>\n",
       "      <td>1553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1757</th>\n",
       "      <td>01February_2011_Tuesday_tagesschau_default-3</td>\n",
       "      <td>149</td>\n",
       "      <td>180</td>\n",
       "      <td>0.931136</td>\n",
       "      <td>4.986779</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1758</th>\n",
       "      <td>15July_2011_Friday_tagesschau_default-2</td>\n",
       "      <td>85</td>\n",
       "      <td>106</td>\n",
       "      <td>0.931136</td>\n",
       "      <td>4.986779</td>\n",
       "      <td>1127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1759</th>\n",
       "      <td>01February_2011_Tuesday_tagesschau_default-3</td>\n",
       "      <td>149</td>\n",
       "      <td>181</td>\n",
       "      <td>0.910373</td>\n",
       "      <td>12.089435</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1760</th>\n",
       "      <td>03February_2011_Thursday_tagesschau_default-11</td>\n",
       "      <td>62</td>\n",
       "      <td>94</td>\n",
       "      <td>0.910373</td>\n",
       "      <td>12.089435</td>\n",
       "      <td>148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1761</th>\n",
       "      <td>01February_2011_Tuesday_tagesschau_default-3</td>\n",
       "      <td>152</td>\n",
       "      <td>186</td>\n",
       "      <td>0.914456</td>\n",
       "      <td>15.447130</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1762</th>\n",
       "      <td>27July_2010_Tuesday_tagesschau_default-5</td>\n",
       "      <td>41</td>\n",
       "      <td>67</td>\n",
       "      <td>0.914456</td>\n",
       "      <td>15.447130</td>\n",
       "      <td>1854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1763</th>\n",
       "      <td>28January_2013_Monday_heute_default-20</td>\n",
       "      <td>42</td>\n",
       "      <td>64</td>\n",
       "      <td>0.918128</td>\n",
       "      <td>12.806324</td>\n",
       "      <td>1895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1764</th>\n",
       "      <td>30May_2011_Monday_heute_default-0</td>\n",
       "      <td>114</td>\n",
       "      <td>138</td>\n",
       "      <td>0.918128</td>\n",
       "      <td>12.806324</td>\n",
       "      <td>2093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1765</th>\n",
       "      <td>28January_2013_Monday_heute_default-20</td>\n",
       "      <td>48</td>\n",
       "      <td>66</td>\n",
       "      <td>0.923482</td>\n",
       "      <td>18.969688</td>\n",
       "      <td>1895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1766</th>\n",
       "      <td>25August_2009_Tuesday_heute_default-5</td>\n",
       "      <td>63</td>\n",
       "      <td>88</td>\n",
       "      <td>0.923482</td>\n",
       "      <td>18.969688</td>\n",
       "      <td>1673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1767</th>\n",
       "      <td>28January_2013_Monday_heute_default-20</td>\n",
       "      <td>19</td>\n",
       "      <td>39</td>\n",
       "      <td>0.904932</td>\n",
       "      <td>4.986779</td>\n",
       "      <td>1895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1768</th>\n",
       "      <td>22September_2010_Wednesday_heute_default-6</td>\n",
       "      <td>68</td>\n",
       "      <td>89</td>\n",
       "      <td>0.904932</td>\n",
       "      <td>4.986779</td>\n",
       "      <td>1476</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1768 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            filename  start  end     score  \\\n",
       "1          26June_2010_Saturday_tagesschau_default-2     72   98  0.921689   \n",
       "2         03July_2010_Saturday_tagesschau_default-14    103  128  0.921689   \n",
       "3          26June_2010_Saturday_tagesschau_default-2     74   98  0.915857   \n",
       "4         25October_2010_Monday_tagesschau_default-4    158  184  0.915857   \n",
       "5          26June_2010_Saturday_tagesschau_default-2     78   98  0.924372   \n",
       "6              10August_2009_Monday_heute_default-11    101  123  0.924372   \n",
       "7          26June_2010_Saturday_tagesschau_default-0     29   66  0.920092   \n",
       "8      05December_2009_Saturday_tagesschau_default-0     36   70  0.920092   \n",
       "9          26June_2010_Saturday_tagesschau_default-0     18   50  0.903186   \n",
       "10         09August_2009_Sunday_tagesschau_default-0     14   41  0.903186   \n",
       "11         26June_2010_Saturday_tagesschau_default-0     17   54  0.914556   \n",
       "12           25June_2010_Friday_tagesschau_default-0     10   45  0.914556   \n",
       "13         26June_2010_Saturday_tagesschau_default-0     16   46  0.913693   \n",
       "14         17June_2010_Thursday_tagesschau_default-0     11   38  0.913693   \n",
       "15         26June_2010_Saturday_tagesschau_default-0     17   49  0.914637   \n",
       "16        25October_2010_Monday_tagesschau_default-0     14   46  0.914637   \n",
       "17         26June_2010_Saturday_tagesschau_default-0     38   64  0.903906   \n",
       "18      30January_2010_Saturday_tagesschau_default-0     38   68  0.903906   \n",
       "19         26June_2010_Saturday_tagesschau_default-0     19   58  0.924526   \n",
       "20      29January_2011_Saturday_tagesschau_default-0     24   60  0.924526   \n",
       "21           23February_2010_Tuesday_heute_default-1     42   69  0.934854   \n",
       "22            22November_2010_Monday_heute_default-1     61   95  0.934854   \n",
       "23           23February_2010_Tuesday_heute_default-1     40   70  0.930481   \n",
       "24        01June_2011_Wednesday_tagesschau_default-6     40   73  0.930481   \n",
       "25           23February_2010_Tuesday_heute_default-1     41   71  0.912161   \n",
       "26                30May_2011_Monday_heute_default-13     67  106  0.912161   \n",
       "27           23February_2010_Tuesday_heute_default-1     41   73  0.915484   \n",
       "28          12April_2010_Monday_tagesschau_default-2     57   85  0.915484   \n",
       "29           23February_2010_Tuesday_heute_default-1     41   70  0.941494   \n",
       "30            05December_2011_Monday_heute_default-4     64   82  0.941494   \n",
       "...                                              ...    ...  ...       ...   \n",
       "1739    23February_2010_Tuesday_tagesschau_default-3     43   67  0.912950   \n",
       "1740       05October_2011_Wednesday_heute_default-17     31   65  0.912950   \n",
       "1741             31May_2011_Tuesday_heute_default-24      4   32  0.916441   \n",
       "1742        24February_2011_Thursday_heute_default-7     13   42  0.916441   \n",
       "1743             31May_2011_Tuesday_heute_default-24      5   32  0.913738   \n",
       "1744        24February_2011_Thursday_heute_default-7     14   52  0.913738   \n",
       "1745             31May_2011_Tuesday_heute_default-24      9   33  0.908595   \n",
       "1746        24February_2011_Thursday_heute_default-7     26   59  0.908595   \n",
       "1747             31May_2011_Tuesday_heute_default-24      8   32  0.904950   \n",
       "1748   23February_2010_Tuesday_tagesschau_default-12     95  113  0.904950   \n",
       "1749      24August_2010_Tuesday_tagesschau_default-3     73   95  0.905229   \n",
       "1750    01February_2011_Tuesday_tagesschau_default-3    152  183  0.905229   \n",
       "1751      24August_2010_Tuesday_tagesschau_default-3     69   93  0.932088   \n",
       "1752          23May_2011_Monday_tagesschau_default-2    124  152  0.932088   \n",
       "1753      24August_2010_Tuesday_tagesschau_default-3     63   94  0.907828   \n",
       "1754  03February_2011_Thursday_tagesschau_default-11     62   93  0.907828   \n",
       "1755    01February_2011_Tuesday_tagesschau_default-3    152  184  0.900610   \n",
       "1756          23May_2011_Monday_tagesschau_default-2    126  152  0.900610   \n",
       "1757    01February_2011_Tuesday_tagesschau_default-3    149  180  0.931136   \n",
       "1758         15July_2011_Friday_tagesschau_default-2     85  106  0.931136   \n",
       "1759    01February_2011_Tuesday_tagesschau_default-3    149  181  0.910373   \n",
       "1760  03February_2011_Thursday_tagesschau_default-11     62   94  0.910373   \n",
       "1761    01February_2011_Tuesday_tagesschau_default-3    152  186  0.914456   \n",
       "1762        27July_2010_Tuesday_tagesschau_default-5     41   67  0.914456   \n",
       "1763          28January_2013_Monday_heute_default-20     42   64  0.918128   \n",
       "1764               30May_2011_Monday_heute_default-0    114  138  0.918128   \n",
       "1765          28January_2013_Monday_heute_default-20     48   66  0.923482   \n",
       "1766           25August_2009_Tuesday_heute_default-5     63   88  0.923482   \n",
       "1767          28January_2013_Monday_heute_default-20     19   39  0.904932   \n",
       "1768      22September_2010_Wednesday_heute_default-6     68   89  0.904932   \n",
       "\n",
       "        unknown  file_id  \n",
       "1      4.670115     1775  \n",
       "2      4.670115      156  \n",
       "3      3.789952     1775  \n",
       "4      3.789952     1726  \n",
       "5      3.391010     1775  \n",
       "6      3.391010      685  \n",
       "7     23.872665     1769  \n",
       "8     23.872665      240  \n",
       "9      8.026211     1769  \n",
       "10     8.026211      589  \n",
       "11    16.556871     1769  \n",
       "12    16.556871     1701  \n",
       "13     3.632980     1769  \n",
       "14     3.632980     1209  \n",
       "15     5.679434     1769  \n",
       "16     5.679434     1720  \n",
       "17     6.125254     1769  \n",
       "18     6.125254     2043  \n",
       "19    33.653538     1769  \n",
       "20    33.653538     1978  \n",
       "21    39.421871     1504  \n",
       "22    39.421871     1465  \n",
       "23    33.443237     1504  \n",
       "24    33.443237       67  \n",
       "25    37.533344     1504  \n",
       "26    37.533344     2096  \n",
       "27    29.277313     1504  \n",
       "28    29.277313      860  \n",
       "29    21.379238     1504  \n",
       "30    21.379238      248  \n",
       "...         ...      ...  \n",
       "1739  14.888820     1519  \n",
       "1740  14.888820      379  \n",
       "1741  11.166643     2163  \n",
       "1742  11.166643     1607  \n",
       "1743  23.135624     2163  \n",
       "1744  23.135624     1607  \n",
       "1745  12.309090     2163  \n",
       "1746  12.309090     1607  \n",
       "1747   4.587836     2163  \n",
       "1748   4.587836     1515  \n",
       "1749   4.708822     1597  \n",
       "1750   4.708822       23  \n",
       "1751  18.755789     1597  \n",
       "1752  18.755789     1553  \n",
       "1753  11.890221     1597  \n",
       "1754  11.890221      148  \n",
       "1755  26.614849       23  \n",
       "1756  26.614849     1553  \n",
       "1757   4.986779       23  \n",
       "1758   4.986779     1127  \n",
       "1759  12.089435       23  \n",
       "1760  12.089435      148  \n",
       "1761  15.447130       23  \n",
       "1762  15.447130     1854  \n",
       "1763  12.806324     1895  \n",
       "1764  12.806324     2093  \n",
       "1765  18.969688     1895  \n",
       "1766  18.969688     1673  \n",
       "1767   4.986779     1895  \n",
       "1768   4.986779     1476  \n",
       "\n",
       "[1768 rows x 6 columns]"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nodes_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils.db_utils import nodes_with_info, gold_fragments_df_for_signer, nodes_with_types\n",
    "\n",
    "clusters_set = set()\n",
    "for row in clusters_list: clusters_set |= set(row)\n",
    "all_centroids = list(clusters_set)\n",
    "\n",
    "\n",
    "# select only nodes that belong to a cluster (dedups?)\n",
    "nodes_df = nodes_df.loc[all_centroids]\n",
    "\n",
    "gold_fragments = gold_fragments_df_for_signer(signer_id)\n",
    "\n",
    "nodes_df = nodes_with_info(nodes_df=nodes_df, gold_fragments=gold_fragments)\n",
    "\n",
    "nodes_df = nodes_with_types(nodes_df=nodes_df, gold_fragments=gold_fragments)\n",
    "\n",
    "for clus in clusters_list:\n",
    "    n_signer_different = len(nodes_df.signer_id[clus].unique())\n",
    "    if n_signer_different>1:\n",
    "        print(clus,n_signer_different)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'clusters_list' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-499720b03ac2>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mfor\u001b[0m \u001b[0mclus\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclusters_list\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m     \u001b[0mn_signer_different\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnodes_df\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msigner_id\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclus\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0munique\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mn_signer_different\u001b[0m\u001b[0;34m>\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m         \u001b[0;32mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mn_signer_different\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'clusters_list' is not defined"
     ]
    }
   ],
   "source": [
    "for clus in clusters_list:\n",
    "    n_signer_different = len(nodes_df.signer_id[clus].unique())\n",
    "    if n_signer_different>1:\n",
    "        print(clus,n_signer_different)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filename</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>score</th>\n",
       "      <th>unknown</th>\n",
       "      <th>file_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>24576</th>\n",
       "      <td>02June_2010_Wednesday_tagesschau_default-3</td>\n",
       "      <td>3</td>\n",
       "      <td>39</td>\n",
       "      <td>0.914838</td>\n",
       "      <td>28.765520</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32769</th>\n",
       "      <td>30May_2010_Sunday_tagesschau_default-0</td>\n",
       "      <td>40</td>\n",
       "      <td>106</td>\n",
       "      <td>0.960022</td>\n",
       "      <td>53.500641</td>\n",
       "      <td>2085</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32770</th>\n",
       "      <td>05July_2010_Monday_tagesschau_default-19</td>\n",
       "      <td>2</td>\n",
       "      <td>71</td>\n",
       "      <td>0.960022</td>\n",
       "      <td>53.500641</td>\n",
       "      <td>321</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24579</th>\n",
       "      <td>26October_2010_Tuesday_tagesschau_default-8</td>\n",
       "      <td>14</td>\n",
       "      <td>44</td>\n",
       "      <td>0.922156</td>\n",
       "      <td>14.963890</td>\n",
       "      <td>1796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24580</th>\n",
       "      <td>20March_2010_Saturday_heute_default-10</td>\n",
       "      <td>47</td>\n",
       "      <td>85</td>\n",
       "      <td>0.922156</td>\n",
       "      <td>14.963890</td>\n",
       "      <td>1341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57345</th>\n",
       "      <td>24June_2010_Thursday_tagesschau_default-17</td>\n",
       "      <td>105</td>\n",
       "      <td>135</td>\n",
       "      <td>0.917187</td>\n",
       "      <td>8.719263</td>\n",
       "      <td>1643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90120</th>\n",
       "      <td>11August_2011_Thursday_heute_default-6</td>\n",
       "      <td>39</td>\n",
       "      <td>66</td>\n",
       "      <td>0.915549</td>\n",
       "      <td>7.518915</td>\n",
       "      <td>780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57355</th>\n",
       "      <td>24June_2010_Thursday_tagesschau_default-17</td>\n",
       "      <td>52</td>\n",
       "      <td>78</td>\n",
       "      <td>0.912963</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>1643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57356</th>\n",
       "      <td>06May_2011_Friday_tagesschau_default-16</td>\n",
       "      <td>26</td>\n",
       "      <td>51</td>\n",
       "      <td>0.912963</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57346</th>\n",
       "      <td>07August_2009_Friday_tagesschau_default-13</td>\n",
       "      <td>12</td>\n",
       "      <td>28</td>\n",
       "      <td>0.917187</td>\n",
       "      <td>8.719263</td>\n",
       "      <td>487</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81935</th>\n",
       "      <td>04April_2011_Monday_heute_default-11</td>\n",
       "      <td>93</td>\n",
       "      <td>111</td>\n",
       "      <td>0.917394</td>\n",
       "      <td>4.259697</td>\n",
       "      <td>165</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81936</th>\n",
       "      <td>06October_2011_Thursday_heute_default-9</td>\n",
       "      <td>83</td>\n",
       "      <td>111</td>\n",
       "      <td>0.917394</td>\n",
       "      <td>4.259697</td>\n",
       "      <td>456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40983</th>\n",
       "      <td>24June_2010_Thursday_tagesschau_default-2</td>\n",
       "      <td>53</td>\n",
       "      <td>79</td>\n",
       "      <td>0.917052</td>\n",
       "      <td>10.419792</td>\n",
       "      <td>1644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40984</th>\n",
       "      <td>24June_2010_Thursday_tagesschau_default-2</td>\n",
       "      <td>109</td>\n",
       "      <td>135</td>\n",
       "      <td>0.917052</td>\n",
       "      <td>10.419792</td>\n",
       "      <td>1644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24601</th>\n",
       "      <td>09November_2010_Tuesday_heute_default-6</td>\n",
       "      <td>45</td>\n",
       "      <td>61</td>\n",
       "      <td>0.922648</td>\n",
       "      <td>4.456918</td>\n",
       "      <td>664</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24602</th>\n",
       "      <td>27July_2010_Tuesday_tagesschau_default-3</td>\n",
       "      <td>94</td>\n",
       "      <td>111</td>\n",
       "      <td>0.922648</td>\n",
       "      <td>4.456918</td>\n",
       "      <td>1853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24603</th>\n",
       "      <td>05December_2009_Saturday_tagesschau_default-0</td>\n",
       "      <td>34</td>\n",
       "      <td>51</td>\n",
       "      <td>0.911625</td>\n",
       "      <td>19.554518</td>\n",
       "      <td>240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24604</th>\n",
       "      <td>30November_2010_Tuesday_tagesschau_default-0</td>\n",
       "      <td>64</td>\n",
       "      <td>90</td>\n",
       "      <td>0.911625</td>\n",
       "      <td>19.554518</td>\n",
       "      <td>2105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24605</th>\n",
       "      <td>14July_2011_Thursday_heute_default-6</td>\n",
       "      <td>9</td>\n",
       "      <td>29</td>\n",
       "      <td>0.909970</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>1063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24606</th>\n",
       "      <td>06January_2011_Thursday_tagesschau_default-7</td>\n",
       "      <td>38</td>\n",
       "      <td>48</td>\n",
       "      <td>0.909970</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24609</th>\n",
       "      <td>28January_2013_Monday_heute_default-20</td>\n",
       "      <td>120</td>\n",
       "      <td>133</td>\n",
       "      <td>0.911445</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>1895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24610</th>\n",
       "      <td>05January_2011_Wednesday_heute_default-15</td>\n",
       "      <td>116</td>\n",
       "      <td>134</td>\n",
       "      <td>0.911445</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>287</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43697</th>\n",
       "      <td>29January_2011_Saturday_tagesschau_default-12</td>\n",
       "      <td>37</td>\n",
       "      <td>55</td>\n",
       "      <td>0.920139</td>\n",
       "      <td>13.979062</td>\n",
       "      <td>1980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24617</th>\n",
       "      <td>21March_2011_Monday_tagesschau_default-13</td>\n",
       "      <td>34</td>\n",
       "      <td>56</td>\n",
       "      <td>0.911704</td>\n",
       "      <td>16.540947</td>\n",
       "      <td>1403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24618</th>\n",
       "      <td>11November_2010_Thursday_tagesschau_default-4</td>\n",
       "      <td>76</td>\n",
       "      <td>96</td>\n",
       "      <td>0.911704</td>\n",
       "      <td>16.540947</td>\n",
       "      <td>841</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90119</th>\n",
       "      <td>23May_2011_Monday_heute_default-0</td>\n",
       "      <td>105</td>\n",
       "      <td>119</td>\n",
       "      <td>0.915549</td>\n",
       "      <td>7.518915</td>\n",
       "      <td>1538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24620</th>\n",
       "      <td>26April_2010_Monday_tagesschau_default-15</td>\n",
       "      <td>71</td>\n",
       "      <td>93</td>\n",
       "      <td>0.917024</td>\n",
       "      <td>18.894093</td>\n",
       "      <td>1738</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68274</th>\n",
       "      <td>09November_2010_Tuesday_heute_default-2</td>\n",
       "      <td>19</td>\n",
       "      <td>42</td>\n",
       "      <td>0.915430</td>\n",
       "      <td>17.560314</td>\n",
       "      <td>661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81966</th>\n",
       "      <td>21November_2010_Sunday_tagesschau_default-2</td>\n",
       "      <td>30</td>\n",
       "      <td>56</td>\n",
       "      <td>0.912031</td>\n",
       "      <td>11.547912</td>\n",
       "      <td>1414</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81967</th>\n",
       "      <td>05February_2010_Friday_tagesschau_default-2</td>\n",
       "      <td>39</td>\n",
       "      <td>75</td>\n",
       "      <td>0.910343</td>\n",
       "      <td>11.547912</td>\n",
       "      <td>271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24526</th>\n",
       "      <td>10March_2011_Thursday_heute_default-7</td>\n",
       "      <td>114</td>\n",
       "      <td>145</td>\n",
       "      <td>0.915909</td>\n",
       "      <td>6.418840</td>\n",
       "      <td>728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24527</th>\n",
       "      <td>26March_2011_Saturday_tagesschau_default-8</td>\n",
       "      <td>159</td>\n",
       "      <td>177</td>\n",
       "      <td>0.924826</td>\n",
       "      <td>3.368059</td>\n",
       "      <td>1785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24528</th>\n",
       "      <td>25February_2011_Friday_tagesschau_default-1</td>\n",
       "      <td>124</td>\n",
       "      <td>150</td>\n",
       "      <td>0.924826</td>\n",
       "      <td>3.368059</td>\n",
       "      <td>1680</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24529</th>\n",
       "      <td>26March_2011_Saturday_tagesschau_default-8</td>\n",
       "      <td>158</td>\n",
       "      <td>171</td>\n",
       "      <td>0.924030</td>\n",
       "      <td>2.992067</td>\n",
       "      <td>1785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24530</th>\n",
       "      <td>05May_2011_Thursday_heute_default-19</td>\n",
       "      <td>27</td>\n",
       "      <td>45</td>\n",
       "      <td>0.924030</td>\n",
       "      <td>2.992067</td>\n",
       "      <td>338</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68259</th>\n",
       "      <td>04May_2011_Wednesday_heute_default-3</td>\n",
       "      <td>118</td>\n",
       "      <td>136</td>\n",
       "      <td>0.910871</td>\n",
       "      <td>13.948633</td>\n",
       "      <td>199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24532</th>\n",
       "      <td>06January_2011_Thursday_tagesschau_default-6</td>\n",
       "      <td>59</td>\n",
       "      <td>80</td>\n",
       "      <td>0.915154</td>\n",
       "      <td>11.958842</td>\n",
       "      <td>419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78287</th>\n",
       "      <td>02June_2010_Wednesday_heute_default-5</td>\n",
       "      <td>16</td>\n",
       "      <td>38</td>\n",
       "      <td>0.916307</td>\n",
       "      <td>2.308117</td>\n",
       "      <td>91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63284</th>\n",
       "      <td>15February_2010_Monday_tagesschau_default-2</td>\n",
       "      <td>17</td>\n",
       "      <td>48</td>\n",
       "      <td>0.916964</td>\n",
       "      <td>5.495231</td>\n",
       "      <td>1081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81880</th>\n",
       "      <td>18January_2010_Monday_tagesschau_default-7</td>\n",
       "      <td>152</td>\n",
       "      <td>182</td>\n",
       "      <td>0.910007</td>\n",
       "      <td>13.691826</td>\n",
       "      <td>1269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24537</th>\n",
       "      <td>26March_2011_Saturday_tagesschau_default-8</td>\n",
       "      <td>8</td>\n",
       "      <td>19</td>\n",
       "      <td>0.930228</td>\n",
       "      <td>16.632748</td>\n",
       "      <td>1785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24538</th>\n",
       "      <td>07May_2010_Friday_tagesschau_default-15</td>\n",
       "      <td>50</td>\n",
       "      <td>75</td>\n",
       "      <td>0.930228</td>\n",
       "      <td>16.632748</td>\n",
       "      <td>523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24539</th>\n",
       "      <td>26March_2011_Saturday_tagesschau_default-8</td>\n",
       "      <td>154</td>\n",
       "      <td>177</td>\n",
       "      <td>0.909943</td>\n",
       "      <td>10.921698</td>\n",
       "      <td>1785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24540</th>\n",
       "      <td>23May_2011_Monday_tagesschau_default-12</td>\n",
       "      <td>142</td>\n",
       "      <td>171</td>\n",
       "      <td>0.909943</td>\n",
       "      <td>10.921698</td>\n",
       "      <td>1550</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32187</th>\n",
       "      <td>23September_2010_Thursday_tagesschau_default-12</td>\n",
       "      <td>117</td>\n",
       "      <td>136</td>\n",
       "      <td>0.913668</td>\n",
       "      <td>7.147704</td>\n",
       "      <td>1574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51009</th>\n",
       "      <td>06December_2009_Sunday_tagesschau_default-15</td>\n",
       "      <td>2</td>\n",
       "      <td>18</td>\n",
       "      <td>0.909985</td>\n",
       "      <td>3.168588</td>\n",
       "      <td>389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24547</th>\n",
       "      <td>26March_2011_Saturday_tagesschau_default-8</td>\n",
       "      <td>159</td>\n",
       "      <td>173</td>\n",
       "      <td>0.917979</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>1785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24548</th>\n",
       "      <td>23May_2011_Monday_heute_default-6</td>\n",
       "      <td>66</td>\n",
       "      <td>88</td>\n",
       "      <td>0.917979</td>\n",
       "      <td>2.194183</td>\n",
       "      <td>1544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24555</th>\n",
       "      <td>28February_2011_Monday_tagesschau_default-7</td>\n",
       "      <td>143</td>\n",
       "      <td>183</td>\n",
       "      <td>0.916400</td>\n",
       "      <td>27.230654</td>\n",
       "      <td>1887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24556</th>\n",
       "      <td>09March_2011_Wednesday_heute_default-11</td>\n",
       "      <td>8</td>\n",
       "      <td>37</td>\n",
       "      <td>0.916400</td>\n",
       "      <td>27.230654</td>\n",
       "      <td>641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49864</th>\n",
       "      <td>22December_2010_Wednesday_tagesschau_default-16</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>0.921303</td>\n",
       "      <td>2.593125</td>\n",
       "      <td>1432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49137</th>\n",
       "      <td>22November_2010_Monday_heute_default-7</td>\n",
       "      <td>33</td>\n",
       "      <td>66</td>\n",
       "      <td>0.916693</td>\n",
       "      <td>125.270355</td>\n",
       "      <td>1469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49138</th>\n",
       "      <td>28October_2010_Thursday_heute_default-2</td>\n",
       "      <td>3</td>\n",
       "      <td>29</td>\n",
       "      <td>0.916693</td>\n",
       "      <td>125.270355</td>\n",
       "      <td>1961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24563</th>\n",
       "      <td>28February_2011_Monday_tagesschau_default-7</td>\n",
       "      <td>60</td>\n",
       "      <td>74</td>\n",
       "      <td>0.911022</td>\n",
       "      <td>10.422581</td>\n",
       "      <td>1887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24564</th>\n",
       "      <td>10January_2011_Monday_heute_default-5</td>\n",
       "      <td>146</td>\n",
       "      <td>166</td>\n",
       "      <td>0.911022</td>\n",
       "      <td>10.422581</td>\n",
       "      <td>707</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98294</th>\n",
       "      <td>12July_2010_Monday_tagesschau_default-0</td>\n",
       "      <td>35</td>\n",
       "      <td>62</td>\n",
       "      <td>0.927679</td>\n",
       "      <td>22.295851</td>\n",
       "      <td>913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98295</th>\n",
       "      <td>08November_2010_Monday_heute_default-9</td>\n",
       "      <td>22</td>\n",
       "      <td>50</td>\n",
       "      <td>0.911207</td>\n",
       "      <td>6.545053</td>\n",
       "      <td>569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32188</th>\n",
       "      <td>24February_2011_Thursday_heute_default-4</td>\n",
       "      <td>73</td>\n",
       "      <td>95</td>\n",
       "      <td>0.913668</td>\n",
       "      <td>7.147704</td>\n",
       "      <td>1604</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86199</th>\n",
       "      <td>26April_2010_Monday_tagesschau_default-0</td>\n",
       "      <td>78</td>\n",
       "      <td>93</td>\n",
       "      <td>0.922675</td>\n",
       "      <td>6.372089</td>\n",
       "      <td>1735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49151</th>\n",
       "      <td>22November_2010_Monday_heute_default-7</td>\n",
       "      <td>12</td>\n",
       "      <td>33</td>\n",
       "      <td>0.929510</td>\n",
       "      <td>53.964272</td>\n",
       "      <td>1469</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3742 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              filename  start  end     score  \\\n",
       "24576       02June_2010_Wednesday_tagesschau_default-3      3   39  0.914838   \n",
       "32769           30May_2010_Sunday_tagesschau_default-0     40  106  0.960022   \n",
       "32770         05July_2010_Monday_tagesschau_default-19      2   71  0.960022   \n",
       "24579      26October_2010_Tuesday_tagesschau_default-8     14   44  0.922156   \n",
       "24580           20March_2010_Saturday_heute_default-10     47   85  0.922156   \n",
       "57345       24June_2010_Thursday_tagesschau_default-17    105  135  0.917187   \n",
       "90120           11August_2011_Thursday_heute_default-6     39   66  0.915549   \n",
       "57355       24June_2010_Thursday_tagesschau_default-17     52   78  0.912963   \n",
       "57356          06May_2011_Friday_tagesschau_default-16     26   51  0.912963   \n",
       "57346       07August_2009_Friday_tagesschau_default-13     12   28  0.917187   \n",
       "81935             04April_2011_Monday_heute_default-11     93  111  0.917394   \n",
       "81936          06October_2011_Thursday_heute_default-9     83  111  0.917394   \n",
       "40983        24June_2010_Thursday_tagesschau_default-2     53   79  0.917052   \n",
       "40984        24June_2010_Thursday_tagesschau_default-2    109  135  0.917052   \n",
       "24601          09November_2010_Tuesday_heute_default-6     45   61  0.922648   \n",
       "24602         27July_2010_Tuesday_tagesschau_default-3     94  111  0.922648   \n",
       "24603    05December_2009_Saturday_tagesschau_default-0     34   51  0.911625   \n",
       "24604     30November_2010_Tuesday_tagesschau_default-0     64   90  0.911625   \n",
       "24605             14July_2011_Thursday_heute_default-6      9   29  0.909970   \n",
       "24606     06January_2011_Thursday_tagesschau_default-7     38   48  0.909970   \n",
       "24609           28January_2013_Monday_heute_default-20    120  133  0.911445   \n",
       "24610        05January_2011_Wednesday_heute_default-15    116  134  0.911445   \n",
       "43697    29January_2011_Saturday_tagesschau_default-12     37   55  0.920139   \n",
       "24617        21March_2011_Monday_tagesschau_default-13     34   56  0.911704   \n",
       "24618    11November_2010_Thursday_tagesschau_default-4     76   96  0.911704   \n",
       "90119                23May_2011_Monday_heute_default-0    105  119  0.915549   \n",
       "24620        26April_2010_Monday_tagesschau_default-15     71   93  0.917024   \n",
       "68274          09November_2010_Tuesday_heute_default-2     19   42  0.915430   \n",
       "81966      21November_2010_Sunday_tagesschau_default-2     30   56  0.912031   \n",
       "81967      05February_2010_Friday_tagesschau_default-2     39   75  0.910343   \n",
       "...                                                ...    ...  ...       ...   \n",
       "24526            10March_2011_Thursday_heute_default-7    114  145  0.915909   \n",
       "24527       26March_2011_Saturday_tagesschau_default-8    159  177  0.924826   \n",
       "24528      25February_2011_Friday_tagesschau_default-1    124  150  0.924826   \n",
       "24529       26March_2011_Saturday_tagesschau_default-8    158  171  0.924030   \n",
       "24530             05May_2011_Thursday_heute_default-19     27   45  0.924030   \n",
       "68259             04May_2011_Wednesday_heute_default-3    118  136  0.910871   \n",
       "24532     06January_2011_Thursday_tagesschau_default-6     59   80  0.915154   \n",
       "78287            02June_2010_Wednesday_heute_default-5     16   38  0.916307   \n",
       "63284      15February_2010_Monday_tagesschau_default-2     17   48  0.916964   \n",
       "81880       18January_2010_Monday_tagesschau_default-7    152  182  0.910007   \n",
       "24537       26March_2011_Saturday_tagesschau_default-8      8   19  0.930228   \n",
       "24538          07May_2010_Friday_tagesschau_default-15     50   75  0.930228   \n",
       "24539       26March_2011_Saturday_tagesschau_default-8    154  177  0.909943   \n",
       "24540          23May_2011_Monday_tagesschau_default-12    142  171  0.909943   \n",
       "32187  23September_2010_Thursday_tagesschau_default-12    117  136  0.913668   \n",
       "51009     06December_2009_Sunday_tagesschau_default-15      2   18  0.909985   \n",
       "24547       26March_2011_Saturday_tagesschau_default-8    159  173  0.917979   \n",
       "24548                23May_2011_Monday_heute_default-6     66   88  0.917979   \n",
       "24555      28February_2011_Monday_tagesschau_default-7    143  183  0.916400   \n",
       "24556          09March_2011_Wednesday_heute_default-11      8   37  0.916400   \n",
       "49864  22December_2010_Wednesday_tagesschau_default-16      1   19  0.921303   \n",
       "49137           22November_2010_Monday_heute_default-7     33   66  0.916693   \n",
       "49138          28October_2010_Thursday_heute_default-2      3   29  0.916693   \n",
       "24563      28February_2011_Monday_tagesschau_default-7     60   74  0.911022   \n",
       "24564            10January_2011_Monday_heute_default-5    146  166  0.911022   \n",
       "98294          12July_2010_Monday_tagesschau_default-0     35   62  0.927679   \n",
       "98295           08November_2010_Monday_heute_default-9     22   50  0.911207   \n",
       "32188         24February_2011_Thursday_heute_default-4     73   95  0.913668   \n",
       "86199         26April_2010_Monday_tagesschau_default-0     78   93  0.922675   \n",
       "49151           22November_2010_Monday_heute_default-7     12   33  0.929510   \n",
       "\n",
       "          unknown  file_id  \n",
       "24576   28.765520      100  \n",
       "32769   53.500641     2085  \n",
       "32770   53.500641      321  \n",
       "24579   14.963890     1796  \n",
       "24580   14.963890     1341  \n",
       "57345    8.719263     1643  \n",
       "90120    7.518915      780  \n",
       "57355    2.194183     1643  \n",
       "57356    2.194183      424  \n",
       "57346    8.719263      487  \n",
       "81935    4.259697      165  \n",
       "81936    4.259697      456  \n",
       "40983   10.419792     1644  \n",
       "40984   10.419792     1644  \n",
       "24601    4.456918      664  \n",
       "24602    4.456918     1853  \n",
       "24603   19.554518      240  \n",
       "24604   19.554518     2105  \n",
       "24605    2.194183     1063  \n",
       "24606    2.194183      420  \n",
       "24609    2.194183     1895  \n",
       "24610    2.194183      287  \n",
       "43697   13.979062     1980  \n",
       "24617   16.540947     1403  \n",
       "24618   16.540947      841  \n",
       "90119    7.518915     1538  \n",
       "24620   18.894093     1738  \n",
       "68274   17.560314      661  \n",
       "81966   11.547912     1414  \n",
       "81967   11.547912      271  \n",
       "...           ...      ...  \n",
       "24526    6.418840      728  \n",
       "24527    3.368059     1785  \n",
       "24528    3.368059     1680  \n",
       "24529    2.992067     1785  \n",
       "24530    2.992067      338  \n",
       "68259   13.948633      199  \n",
       "24532   11.958842      419  \n",
       "78287    2.308117       91  \n",
       "63284    5.495231     1081  \n",
       "81880   13.691826     1269  \n",
       "24537   16.632748     1785  \n",
       "24538   16.632748      523  \n",
       "24539   10.921698     1785  \n",
       "24540   10.921698     1550  \n",
       "32187    7.147704     1574  \n",
       "51009    3.168588      389  \n",
       "24547    2.194183     1785  \n",
       "24548    2.194183     1544  \n",
       "24555   27.230654     1887  \n",
       "24556   27.230654      641  \n",
       "49864    2.593125     1432  \n",
       "49137  125.270355     1469  \n",
       "49138  125.270355     1961  \n",
       "24563   10.422581     1887  \n",
       "24564   10.422581      707  \n",
       "98294   22.295851      913  \n",
       "98295    6.545053      569  \n",
       "32188    7.147704     1604  \n",
       "86199    6.372089     1735  \n",
       "49151   53.964272     1469  \n",
       "\n",
       "[3742 rows x 6 columns]"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nodes_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'nodes' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-91-c48ea7a4c803>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mnodes\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'nodes' is not defined"
     ]
    }
   ],
   "source": [
    "nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer01_Signer0850_10_4_02_7_05_06_7_06_032'"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "change_post_disc_thr(olapthr=0.9, dedupthr=0.2, durthr=5, rhothr=1000, min_edge_w=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 0.9 0.2 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "olap 0.3, dedup 0.5\n",
      "./post_disc zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032.lsh64 0.9\n",
      "Read 46404 nodes/tokens\n",
      "Read 144 clusters\n",
      "olap 0.3, dedup 0.5\n",
      "./post_disc zrroot0_corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032/exp/corAE_op100_s4s5_64_iter0_self_tanh_Signer0450_10_4_02_7_05_06_7_06_032.lsh64 0.9\n",
      "Read 46404 nodes/tokens\n",
      "Read 449 clusters\n",
      "15.8854289348% OVERLAP\n"
     ]
    }
   ],
   "source": [
    "#results = []\n",
    "for olap in [ 0.3]:\n",
    "    for dedup in [0.5]:\n",
    "        for th in [600, 800]:\n",
    "            change_post_disc_thr(olapthr=olap, dedupthr=dedup, durthr=5, rhothr=1000, min_edge_w=th)\n",
    "\n",
    "            print('olap {}, dedup {}'.format(olap, dedup))\n",
    "\n",
    "            results_path = post_disc(exp_name, dtw)\n",
    "\n",
    "            \" get nodes \"\n",
    "            nodes_df = get_nodes_df(exp_name)\n",
    "\n",
    "            \" get clusters \"\n",
    "            clusters_list = get_clusters_list(exp_name)\n",
    "            if len(clusters_list)==0: continue\n",
    "\n",
    "            scores = evaluate(None, signer_id, matches_df, nodes_df, clusters_list,\n",
    "                              group=3, interp=False, boundary_th=0.5, fast_compute=True)\n",
    "\n",
    "            scores['dtw'] = dtw\n",
    "            scores['exp'] = exp_name\n",
    "            scores['olap'] = olap\n",
    "            scores['dedup'] = dedup\n",
    "            scores['th'] = th\n",
    "\n",
    "            results.append(scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['n_match',\n",
       " 'n_node',\n",
       " 'n_clus',\n",
       " 'mean_len',\n",
       " 'mean_n_type',\n",
       " 'ned',\n",
       " 'coverage',\n",
       " 'matching_P',\n",
       " 'matching_R',\n",
       " 'matching_F',\n",
       " 'purity',\n",
       " 'inv_purity',\n",
       " 'garbage_ratio',\n",
       " 'avg_purity',\n",
       " 'clus_purity',\n",
       " 'clus_purity_inv',\n",
       " 'grouping_P',\n",
       " 'grouping_R',\n",
       " 'grouping_F',\n",
       " 'type_P',\n",
       " 'type_R',\n",
       " 'type_F',\n",
       " 'dtw',\n",
       " 'exp',\n",
       " 'olap',\n",
       " 'dedup']"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results[0].keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean_len</th>\n",
       "      <th>n_node</th>\n",
       "      <th>n_clus</th>\n",
       "      <th>ned</th>\n",
       "      <th>coverage</th>\n",
       "      <th>olap</th>\n",
       "      <th>dedup</th>\n",
       "      <th>th</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>18.012821</td>\n",
       "      <td>156</td>\n",
       "      <td>76</td>\n",
       "      <td>0.809524</td>\n",
       "      <td>0.031197</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.2</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>17.993671</td>\n",
       "      <td>158</td>\n",
       "      <td>76</td>\n",
       "      <td>0.803371</td>\n",
       "      <td>0.031254</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.7</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18.479042</td>\n",
       "      <td>167</td>\n",
       "      <td>76</td>\n",
       "      <td>0.694915</td>\n",
       "      <td>0.031583</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.9</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>22.257898</td>\n",
       "      <td>3672</td>\n",
       "      <td>1665</td>\n",
       "      <td>0.821376</td>\n",
       "      <td>0.490394</td>\n",
       "      <td>0.90</td>\n",
       "      <td>0.2</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>22.271196</td>\n",
       "      <td>3680</td>\n",
       "      <td>1665</td>\n",
       "      <td>0.817960</td>\n",
       "      <td>0.490735</td>\n",
       "      <td>0.90</td>\n",
       "      <td>0.7</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>22.369338</td>\n",
       "      <td>3731</td>\n",
       "      <td>1665</td>\n",
       "      <td>0.792499</td>\n",
       "      <td>0.490860</td>\n",
       "      <td>0.90</td>\n",
       "      <td>0.9</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>25.212441</td>\n",
       "      <td>24162</td>\n",
       "      <td>9906</td>\n",
       "      <td>0.902595</td>\n",
       "      <td>0.718726</td>\n",
       "      <td>0.98</td>\n",
       "      <td>0.2</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>25.221296</td>\n",
       "      <td>24221</td>\n",
       "      <td>9906</td>\n",
       "      <td>0.901972</td>\n",
       "      <td>0.718794</td>\n",
       "      <td>0.98</td>\n",
       "      <td>0.7</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>25.239221</td>\n",
       "      <td>24329</td>\n",
       "      <td>9906</td>\n",
       "      <td>0.898757</td>\n",
       "      <td>0.718794</td>\n",
       "      <td>0.98</td>\n",
       "      <td>0.9</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>18.479042</td>\n",
       "      <td>167</td>\n",
       "      <td>76</td>\n",
       "      <td>0.694915</td>\n",
       "      <td>0.031583</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.9</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>18.479042</td>\n",
       "      <td>167</td>\n",
       "      <td>76</td>\n",
       "      <td>0.694915</td>\n",
       "      <td>0.031583</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.9</td>\n",
       "      <td>200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>20.025210</td>\n",
       "      <td>238</td>\n",
       "      <td>105</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.047227</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.9</td>\n",
       "      <td>500.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>22.699405</td>\n",
       "      <td>1008</td>\n",
       "      <td>449</td>\n",
       "      <td>0.723595</td>\n",
       "      <td>0.204076</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.9</td>\n",
       "      <td>800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>19.324201</td>\n",
       "      <td>219</td>\n",
       "      <td>105</td>\n",
       "      <td>0.766260</td>\n",
       "      <td>0.046421</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.2</td>\n",
       "      <td>500.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>19.473214</td>\n",
       "      <td>224</td>\n",
       "      <td>105</td>\n",
       "      <td>0.748175</td>\n",
       "      <td>0.046784</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.7</td>\n",
       "      <td>500.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>22.297269</td>\n",
       "      <td>952</td>\n",
       "      <td>449</td>\n",
       "      <td>0.787755</td>\n",
       "      <td>0.202849</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.2</td>\n",
       "      <td>800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>22.336820</td>\n",
       "      <td>956</td>\n",
       "      <td>449</td>\n",
       "      <td>0.782137</td>\n",
       "      <td>0.203122</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.7</td>\n",
       "      <td>800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>20.941368</td>\n",
       "      <td>307</td>\n",
       "      <td>144</td>\n",
       "      <td>0.763369</td>\n",
       "      <td>0.069268</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.2</td>\n",
       "      <td>600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>20.993569</td>\n",
       "      <td>311</td>\n",
       "      <td>144</td>\n",
       "      <td>0.757155</td>\n",
       "      <td>0.069540</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.7</td>\n",
       "      <td>600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>20.941368</td>\n",
       "      <td>307</td>\n",
       "      <td>144</td>\n",
       "      <td>0.763369</td>\n",
       "      <td>0.069268</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.2</td>\n",
       "      <td>600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>22.297269</td>\n",
       "      <td>952</td>\n",
       "      <td>449</td>\n",
       "      <td>0.787755</td>\n",
       "      <td>0.202849</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.2</td>\n",
       "      <td>800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>20.941368</td>\n",
       "      <td>307</td>\n",
       "      <td>144</td>\n",
       "      <td>0.763369</td>\n",
       "      <td>0.069268</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.5</td>\n",
       "      <td>600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>22.297269</td>\n",
       "      <td>952</td>\n",
       "      <td>449</td>\n",
       "      <td>0.787755</td>\n",
       "      <td>0.202849</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.5</td>\n",
       "      <td>800.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     mean_len  n_node  n_clus       ned  coverage  olap  dedup     th\n",
       "0   18.012821     156      76  0.809524  0.031197  0.30    0.2    NaN\n",
       "1   17.993671     158      76  0.803371  0.031254  0.30    0.7    NaN\n",
       "2   18.479042     167      76  0.694915  0.031583  0.30    0.9    NaN\n",
       "3   22.257898    3672    1665  0.821376  0.490394  0.90    0.2    NaN\n",
       "4   22.271196    3680    1665  0.817960  0.490735  0.90    0.7    NaN\n",
       "5   22.369338    3731    1665  0.792499  0.490860  0.90    0.9    NaN\n",
       "6   25.212441   24162    9906  0.902595  0.718726  0.98    0.2    NaN\n",
       "7   25.221296   24221    9906  0.901972  0.718794  0.98    0.7    NaN\n",
       "8   25.239221   24329    9906  0.898757  0.718794  0.98    0.9    NaN\n",
       "9   18.479042     167      76  0.694915  0.031583  0.30    0.9    0.0\n",
       "10  18.479042     167      76  0.694915  0.031583  0.30    0.9  200.0\n",
       "11  20.025210     238     105  0.638889  0.047227  0.30    0.9  500.0\n",
       "12  22.699405    1008     449  0.723595  0.204076  0.30    0.9  800.0\n",
       "13  19.324201     219     105  0.766260  0.046421  0.30    0.2  500.0\n",
       "14  19.473214     224     105  0.748175  0.046784  0.30    0.7  500.0\n",
       "15  22.297269     952     449  0.787755  0.202849  0.30    0.2  800.0\n",
       "16  22.336820     956     449  0.782137  0.203122  0.30    0.7  800.0\n",
       "17  20.941368     307     144  0.763369  0.069268  0.30    0.2  600.0\n",
       "18  20.993569     311     144  0.757155  0.069540  0.30    0.7  600.0\n",
       "19  20.941368     307     144  0.763369  0.069268  0.30    0.2  600.0\n",
       "20  22.297269     952     449  0.787755  0.202849  0.30    0.2  800.0\n",
       "21  20.941368     307     144  0.763369  0.069268  0.30    0.5  600.0\n",
       "22  22.297269     952     449  0.787755  0.202849  0.30    0.5  800.0"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame.from_records(results)[['mean_len','n_node','n_clus','ned','coverage','olap','dedup','th']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "zerospeech",
   "language": "python",
   "name": "zerospeech"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.17"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}