plot_test_results.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import time\n",
    "import scipy.io as sio\n",
    "import scipy.sparse as sp\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import matplotlib as mlp\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "import re\n",
    "numeric_const_pattern = r\"\"\"\n",
    "     [-+]? \n",
    "     (?:\n",
    "         (?: \\d* \\. \\d+ )\n",
    "         |\n",
    "         (?: \\d+ \\.? )\n",
    "     )\n",
    "     (?: [Ee] [+-]? \\d+ ) ?\n",
    "     \"\"\"\n",
    "rx = re.compile(numeric_const_pattern, re.VERBOSE)\n",
    "\n",
    "from IPython.display import clear_output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Minimum Weight Connected Dominating Set (MWCDS) results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>ar_db</th>\n",
       "      <th>ar_gb</th>\n",
       "      <th>mwcds_0</th>\n",
       "      <th>mwcds_i</th>\n",
       "      <th>mwds_0</th>\n",
       "      <th>mwds_i</th>\n",
       "      <th>vvv</th>\n",
       "      <th>m</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BA</td>\n",
       "      <td>200</td>\n",
       "      <td>12</td>\n",
       "      <td>0.281997</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>390.119308</td>\n",
       "      <td>338.092124</td>\n",
       "      <td>1.153885</td>\n",
       "      <td>2.550138</td>\n",
       "      <td>2.942565</td>\n",
       "      <td>107.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>132.577984</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BA</td>\n",
       "      <td>250</td>\n",
       "      <td>11</td>\n",
       "      <td>0.276555</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>596.469000</td>\n",
       "      <td>573.550106</td>\n",
       "      <td>1.039960</td>\n",
       "      <td>2.781050</td>\n",
       "      <td>2.892180</td>\n",
       "      <td>141.0</td>\n",
       "      <td>145.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>206.235064</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BA</td>\n",
       "      <td>150</td>\n",
       "      <td>15</td>\n",
       "      <td>0.166337</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>417.623101</td>\n",
       "      <td>372.809994</td>\n",
       "      <td>1.120204</td>\n",
       "      <td>2.296407</td>\n",
       "      <td>2.572444</td>\n",
       "      <td>96.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>162.344902</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BA</td>\n",
       "      <td>150</td>\n",
       "      <td>21</td>\n",
       "      <td>0.292982</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>218.117986</td>\n",
       "      <td>211.683527</td>\n",
       "      <td>1.030397</td>\n",
       "      <td>4.072156</td>\n",
       "      <td>4.195936</td>\n",
       "      <td>62.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>51.983155</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BA</td>\n",
       "      <td>300</td>\n",
       "      <td>10</td>\n",
       "      <td>0.304991</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>693.979717</td>\n",
       "      <td>590.379322</td>\n",
       "      <td>1.175481</td>\n",
       "      <td>3.227548</td>\n",
       "      <td>3.793921</td>\n",
       "      <td>164.0</td>\n",
       "      <td>181.0</td>\n",
       "      <td>108.0</td>\n",
       "      <td>121.0</td>\n",
       "      <td>182.918842</td>\n",
       "      <td>3.0</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7994</th>\n",
       "      <td>WS</td>\n",
       "      <td>100</td>\n",
       "      <td>21</td>\n",
       "      <td>0.336340</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>28.977632</td>\n",
       "      <td>35.434344</td>\n",
       "      <td>0.817784</td>\n",
       "      <td>3.746417</td>\n",
       "      <td>3.063759</td>\n",
       "      <td>19.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>9.458195</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7995</th>\n",
       "      <td>WS</td>\n",
       "      <td>200</td>\n",
       "      <td>24</td>\n",
       "      <td>0.190733</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>43.788411</td>\n",
       "      <td>44.028235</td>\n",
       "      <td>0.994553</td>\n",
       "      <td>2.619434</td>\n",
       "      <td>2.605166</td>\n",
       "      <td>30.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>16.808301</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7996</th>\n",
       "      <td>WS</td>\n",
       "      <td>100</td>\n",
       "      <td>27</td>\n",
       "      <td>0.232670</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>43.076832</td>\n",
       "      <td>39.307392</td>\n",
       "      <td>1.095896</td>\n",
       "      <td>2.525741</td>\n",
       "      <td>2.767950</td>\n",
       "      <td>15.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>15.562719</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7997</th>\n",
       "      <td>WS</td>\n",
       "      <td>150</td>\n",
       "      <td>24</td>\n",
       "      <td>0.183372</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>38.111246</td>\n",
       "      <td>54.656172</td>\n",
       "      <td>0.697291</td>\n",
       "      <td>2.931486</td>\n",
       "      <td>2.044098</td>\n",
       "      <td>26.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>18.644530</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7998</th>\n",
       "      <td>WS</td>\n",
       "      <td>250</td>\n",
       "      <td>18</td>\n",
       "      <td>0.214285</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>80.995419</td>\n",
       "      <td>85.020880</td>\n",
       "      <td>0.952653</td>\n",
       "      <td>2.709192</td>\n",
       "      <td>2.580921</td>\n",
       "      <td>48.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>31.382370</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7999 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     type  size   k         p  mwds  mwcds         gcn      greedy     ratio  \\\n",
       "0      BA   200  12  0.281997   NaN    NaN  390.119308  338.092124  1.153885   \n",
       "1      BA   250  11  0.276555   NaN    NaN  596.469000  573.550106  1.039960   \n",
       "2      BA   150  15  0.166337   NaN    NaN  417.623101  372.809994  1.120204   \n",
       "3      BA   150  21  0.292982   NaN    NaN  218.117986  211.683527  1.030397   \n",
       "4      BA   300  10  0.304991   NaN    NaN  693.979717  590.379322  1.175481   \n",
       "...   ...   ...  ..       ...   ...    ...         ...         ...       ...   \n",
       "7994   WS   100  21  0.336340   NaN    NaN   28.977632   35.434344  0.817784   \n",
       "7995   WS   200  24  0.190733   NaN    NaN   43.788411   44.028235  0.994553   \n",
       "7996   WS   100  27  0.232670   NaN    NaN   43.076832   39.307392  1.095896   \n",
       "7997   WS   150  24  0.183372   NaN    NaN   38.111246   54.656172  0.697291   \n",
       "7998   WS   250  18  0.214285   NaN    NaN   80.995419   85.020880  0.952653   \n",
       "\n",
       "         ar_db     ar_gb  mwcds_0  mwcds_i  mwds_0  mwds_i         vvv    m  \n",
       "0     2.550138  2.942565    107.0    118.0    69.0    73.0  132.577984  3.0  \n",
       "1     2.781050  2.892180    141.0    145.0    87.0    93.0  206.235064  3.0  \n",
       "2     2.296407  2.572444     96.0    102.0    63.0    71.0  162.344902  2.0  \n",
       "3     4.072156  4.195936     62.0     64.0    42.0    43.0   51.983155  6.0  \n",
       "4     3.227548  3.793921    164.0    181.0   108.0   121.0  182.918842  3.0  \n",
       "...        ...       ...      ...      ...     ...     ...         ...  ...  \n",
       "7994  3.746417  3.063759     19.0     21.0    12.0    19.0    9.458195  7.0  \n",
       "7995  2.619434  2.605166     30.0     41.0    18.0    41.0   16.808301  5.0  \n",
       "7996  2.525741  2.767950     15.0     17.0    10.0    12.0   15.562719  6.0  \n",
       "7997  2.931486  2.044098     26.0     27.0    14.0    25.0   18.644530  4.0  \n",
       "7998  2.709192  2.580921     48.0     62.0    29.0    60.0   31.382370  4.0  \n",
       "\n",
       "[7999 rows x 17 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mwcds_csv_dict = {\n",
    "    \"BA\":'./output/result_GCNCDS_deep_ld1_c32_l5_cheb1_diver2_mpy_dpg_policy_BA_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"GRP\":'./output/result_GCNCDS_deep_ld1_c32_l5_cheb1_diver2_mpy_dpg_policy_GRP_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"ER\":'./output/result_GCNCDS_deep_ld1_c32_l5_cheb1_diver2_mpy_dpg_policy_ER_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"WS\":'./output/result_GCNCDS_deep_ld1_c32_l5_cheb1_diver2_mpy_dpg_policy_WS_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "                   }\n",
    "\n",
    "df_mwcds = pd.DataFrame([])\n",
    "for key in mwcds_csv_dict:\n",
    "    mwcds_csv = mwcds_csv_dict[key]\n",
    "    try:\n",
    "        df_tmp = pd.read_csv(mwcds_csv, index_col=0)\n",
    "        df_tmp['m'] = df_tmp['p']*df_tmp['k']\n",
    "        df_tmp['m'] = df_tmp['m'].apply(np.round)\n",
    "        df_tmp['type']=key\n",
    "        df_mwcds = pd.concat([df_mwcds, df_tmp], axis=0, ignore_index=True)\n",
    "    except:\n",
    "        pass\n",
    "df_mwcds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>ar_db</th>\n",
       "      <th>ar_gb</th>\n",
       "      <th>mwcds_0</th>\n",
       "      <th>mwcds_i</th>\n",
       "      <th>mwds_0</th>\n",
       "      <th>mwds_i</th>\n",
       "      <th>vvv</th>\n",
       "      <th>m</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>type</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BA</th>\n",
       "      <td>225.350000</td>\n",
       "      <td>19.59750</td>\n",
       "      <td>0.250471</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>338.655710</td>\n",
       "      <td>344.825357</td>\n",
       "      <td>0.960508</td>\n",
       "      <td>3.513142</td>\n",
       "      <td>3.308604</td>\n",
       "      <td>99.871000</td>\n",
       "      <td>97.2040</td>\n",
       "      <td>65.245500</td>\n",
       "      <td>64.405000</td>\n",
       "      <td>103.060581</td>\n",
       "      <td>4.893000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ER</th>\n",
       "      <td>226.538269</td>\n",
       "      <td>19.29965</td>\n",
       "      <td>0.251120</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>105.870887</td>\n",
       "      <td>124.065012</td>\n",
       "      <td>0.840088</td>\n",
       "      <td>3.122040</td>\n",
       "      <td>2.531223</td>\n",
       "      <td>53.451726</td>\n",
       "      <td>53.4002</td>\n",
       "      <td>34.144072</td>\n",
       "      <td>43.469735</td>\n",
       "      <td>43.026415</td>\n",
       "      <td>4.848924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GRP</th>\n",
       "      <td>225.675000</td>\n",
       "      <td>19.53200</td>\n",
       "      <td>0.252236</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>64.612016</td>\n",
       "      <td>79.894380</td>\n",
       "      <td>0.827460</td>\n",
       "      <td>3.513383</td>\n",
       "      <td>2.823001</td>\n",
       "      <td>38.965000</td>\n",
       "      <td>44.3905</td>\n",
       "      <td>25.427000</td>\n",
       "      <td>41.031500</td>\n",
       "      <td>24.013723</td>\n",
       "      <td>4.899500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WS</th>\n",
       "      <td>224.625000</td>\n",
       "      <td>19.65800</td>\n",
       "      <td>0.249385</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>82.251777</td>\n",
       "      <td>90.936149</td>\n",
       "      <td>0.934257</td>\n",
       "      <td>2.589077</td>\n",
       "      <td>2.348184</td>\n",
       "      <td>45.057500</td>\n",
       "      <td>50.6340</td>\n",
       "      <td>27.530000</td>\n",
       "      <td>43.666000</td>\n",
       "      <td>39.282312</td>\n",
       "      <td>4.910000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            size         k         p  mwds  mwcds         gcn      greedy  \\\n",
       "type                                                                        \n",
       "BA    225.350000  19.59750  0.250471   NaN    NaN  338.655710  344.825357   \n",
       "ER    226.538269  19.29965  0.251120   NaN    NaN  105.870887  124.065012   \n",
       "GRP   225.675000  19.53200  0.252236   NaN    NaN   64.612016   79.894380   \n",
       "WS    224.625000  19.65800  0.249385   NaN    NaN   82.251777   90.936149   \n",
       "\n",
       "         ratio     ar_db     ar_gb    mwcds_0  mwcds_i     mwds_0     mwds_i  \\\n",
       "type                                                                           \n",
       "BA    0.960508  3.513142  3.308604  99.871000  97.2040  65.245500  64.405000   \n",
       "ER    0.840088  3.122040  2.531223  53.451726  53.4002  34.144072  43.469735   \n",
       "GRP   0.827460  3.513383  2.823001  38.965000  44.3905  25.427000  41.031500   \n",
       "WS    0.934257  2.589077  2.348184  45.057500  50.6340  27.530000  43.666000   \n",
       "\n",
       "             vvv         m  \n",
       "type                        \n",
       "BA    103.060581  4.893000  \n",
       "ER     43.026415  4.848924  \n",
       "GRP    24.013723  4.899500  \n",
       "WS     39.282312  4.910000  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mwcds.groupby(['type']).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>ar_db</th>\n",
       "      <th>ar_gb</th>\n",
       "      <th>mwcds_0</th>\n",
       "      <th>mwcds_i</th>\n",
       "      <th>mwds_0</th>\n",
       "      <th>mwds_i</th>\n",
       "      <th>vvv</th>\n",
       "      <th>m</th>\n",
       "      <th>Solver</th>\n",
       "      <th>ar</th>\n",
       "      <th>Graph</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BA</td>\n",
       "      <td>200</td>\n",
       "      <td>12</td>\n",
       "      <td>0.281997</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>390.119308</td>\n",
       "      <td>338.092124</td>\n",
       "      <td>1.153885</td>\n",
       "      <td>2.550138</td>\n",
       "      <td>2.942565</td>\n",
       "      <td>107.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>132.577984</td>\n",
       "      <td>3.0</td>\n",
       "      <td>GCNN(5)+Dist.Greedy</td>\n",
       "      <td>2.942565</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BA</td>\n",
       "      <td>250</td>\n",
       "      <td>11</td>\n",
       "      <td>0.276555</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>596.469000</td>\n",
       "      <td>573.550106</td>\n",
       "      <td>1.039960</td>\n",
       "      <td>2.781050</td>\n",
       "      <td>2.892180</td>\n",
       "      <td>141.0</td>\n",
       "      <td>145.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>206.235064</td>\n",
       "      <td>3.0</td>\n",
       "      <td>GCNN(5)+Dist.Greedy</td>\n",
       "      <td>2.892180</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BA</td>\n",
       "      <td>150</td>\n",
       "      <td>15</td>\n",
       "      <td>0.166337</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>417.623101</td>\n",
       "      <td>372.809994</td>\n",
       "      <td>1.120204</td>\n",
       "      <td>2.296407</td>\n",
       "      <td>2.572444</td>\n",
       "      <td>96.0</td>\n",
       "      <td>102.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>162.344902</td>\n",
       "      <td>2.0</td>\n",
       "      <td>GCNN(5)+Dist.Greedy</td>\n",
       "      <td>2.572444</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BA</td>\n",
       "      <td>150</td>\n",
       "      <td>21</td>\n",
       "      <td>0.292982</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>218.117986</td>\n",
       "      <td>211.683527</td>\n",
       "      <td>1.030397</td>\n",
       "      <td>4.072156</td>\n",
       "      <td>4.195936</td>\n",
       "      <td>62.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>51.983155</td>\n",
       "      <td>6.0</td>\n",
       "      <td>GCNN(5)+Dist.Greedy</td>\n",
       "      <td>4.195936</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BA</td>\n",
       "      <td>300</td>\n",
       "      <td>10</td>\n",
       "      <td>0.304991</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>693.979717</td>\n",
       "      <td>590.379322</td>\n",
       "      <td>1.175481</td>\n",
       "      <td>3.227548</td>\n",
       "      <td>3.793921</td>\n",
       "      <td>164.0</td>\n",
       "      <td>181.0</td>\n",
       "      <td>108.0</td>\n",
       "      <td>121.0</td>\n",
       "      <td>182.918842</td>\n",
       "      <td>3.0</td>\n",
       "      <td>GCNN(5)+Dist.Greedy</td>\n",
       "      <td>3.793921</td>\n",
       "      <td>BA</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15993</th>\n",
       "      <td>WS</td>\n",
       "      <td>100</td>\n",
       "      <td>21</td>\n",
       "      <td>0.336340</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>28.977632</td>\n",
       "      <td>35.434344</td>\n",
       "      <td>0.817784</td>\n",
       "      <td>3.746417</td>\n",
       "      <td>3.063759</td>\n",
       "      <td>19.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>9.458195</td>\n",
       "      <td>7.0</td>\n",
       "      <td>Dist.Greedy</td>\n",
       "      <td>3.746417</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15994</th>\n",
       "      <td>WS</td>\n",
       "      <td>200</td>\n",
       "      <td>24</td>\n",
       "      <td>0.190733</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>43.788411</td>\n",
       "      <td>44.028235</td>\n",
       "      <td>0.994553</td>\n",
       "      <td>2.619434</td>\n",
       "      <td>2.605166</td>\n",
       "      <td>30.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>16.808301</td>\n",
       "      <td>5.0</td>\n",
       "      <td>Dist.Greedy</td>\n",
       "      <td>2.619434</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15995</th>\n",
       "      <td>WS</td>\n",
       "      <td>100</td>\n",
       "      <td>27</td>\n",
       "      <td>0.232670</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>43.076832</td>\n",
       "      <td>39.307392</td>\n",
       "      <td>1.095896</td>\n",
       "      <td>2.525741</td>\n",
       "      <td>2.767950</td>\n",
       "      <td>15.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>15.562719</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Dist.Greedy</td>\n",
       "      <td>2.525741</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15996</th>\n",
       "      <td>WS</td>\n",
       "      <td>150</td>\n",
       "      <td>24</td>\n",
       "      <td>0.183372</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>38.111246</td>\n",
       "      <td>54.656172</td>\n",
       "      <td>0.697291</td>\n",
       "      <td>2.931486</td>\n",
       "      <td>2.044098</td>\n",
       "      <td>26.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>18.644530</td>\n",
       "      <td>4.0</td>\n",
       "      <td>Dist.Greedy</td>\n",
       "      <td>2.931486</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15997</th>\n",
       "      <td>WS</td>\n",
       "      <td>250</td>\n",
       "      <td>18</td>\n",
       "      <td>0.214285</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>80.995419</td>\n",
       "      <td>85.020880</td>\n",
       "      <td>0.952653</td>\n",
       "      <td>2.709192</td>\n",
       "      <td>2.580921</td>\n",
       "      <td>48.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>31.382370</td>\n",
       "      <td>4.0</td>\n",
       "      <td>Dist.Greedy</td>\n",
       "      <td>2.709192</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>15998 rows × 20 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      type  size   k         p  mwds  mwcds         gcn      greedy     ratio  \\\n",
       "0       BA   200  12  0.281997   NaN    NaN  390.119308  338.092124  1.153885   \n",
       "1       BA   250  11  0.276555   NaN    NaN  596.469000  573.550106  1.039960   \n",
       "2       BA   150  15  0.166337   NaN    NaN  417.623101  372.809994  1.120204   \n",
       "3       BA   150  21  0.292982   NaN    NaN  218.117986  211.683527  1.030397   \n",
       "4       BA   300  10  0.304991   NaN    NaN  693.979717  590.379322  1.175481   \n",
       "...    ...   ...  ..       ...   ...    ...         ...         ...       ...   \n",
       "15993   WS   100  21  0.336340   NaN    NaN   28.977632   35.434344  0.817784   \n",
       "15994   WS   200  24  0.190733   NaN    NaN   43.788411   44.028235  0.994553   \n",
       "15995   WS   100  27  0.232670   NaN    NaN   43.076832   39.307392  1.095896   \n",
       "15996   WS   150  24  0.183372   NaN    NaN   38.111246   54.656172  0.697291   \n",
       "15997   WS   250  18  0.214285   NaN    NaN   80.995419   85.020880  0.952653   \n",
       "\n",
       "          ar_db     ar_gb  mwcds_0  mwcds_i  mwds_0  mwds_i         vvv    m  \\\n",
       "0      2.550138  2.942565    107.0    118.0    69.0    73.0  132.577984  3.0   \n",
       "1      2.781050  2.892180    141.0    145.0    87.0    93.0  206.235064  3.0   \n",
       "2      2.296407  2.572444     96.0    102.0    63.0    71.0  162.344902  2.0   \n",
       "3      4.072156  4.195936     62.0     64.0    42.0    43.0   51.983155  6.0   \n",
       "4      3.227548  3.793921    164.0    181.0   108.0   121.0  182.918842  3.0   \n",
       "...         ...       ...      ...      ...     ...     ...         ...  ...   \n",
       "15993  3.746417  3.063759     19.0     21.0    12.0    19.0    9.458195  7.0   \n",
       "15994  2.619434  2.605166     30.0     41.0    18.0    41.0   16.808301  5.0   \n",
       "15995  2.525741  2.767950     15.0     17.0    10.0    12.0   15.562719  6.0   \n",
       "15996  2.931486  2.044098     26.0     27.0    14.0    25.0   18.644530  4.0   \n",
       "15997  2.709192  2.580921     48.0     62.0    29.0    60.0   31.382370  4.0   \n",
       "\n",
       "                    Solver        ar Graph  \n",
       "0      GCNN(5)+Dist.Greedy  2.942565    BA  \n",
       "1      GCNN(5)+Dist.Greedy  2.892180    BA  \n",
       "2      GCNN(5)+Dist.Greedy  2.572444    BA  \n",
       "3      GCNN(5)+Dist.Greedy  4.195936    BA  \n",
       "4      GCNN(5)+Dist.Greedy  3.793921    BA  \n",
       "...                    ...       ...   ...  \n",
       "15993          Dist.Greedy  3.746417    WS  \n",
       "15994          Dist.Greedy  2.619434    WS  \n",
       "15995          Dist.Greedy  2.525741    WS  \n",
       "15996          Dist.Greedy  2.931486    WS  \n",
       "15997          Dist.Greedy  2.709192    WS  \n",
       "\n",
       "[15998 rows x 20 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = df_mwcds.copy()\n",
    "df2 = df_mwcds.copy()\n",
    "# df1['algo'] = 'GCNN(5)+Dist.Greedy, ' + df1['type']\n",
    "df1['Solver'] = 'GCNN(5)+Dist.Greedy'\n",
    "df1['ar'] = df1['ar_gb']\n",
    "# df2['algo'] = 'Dist.Greedy, ' + df2['type']\n",
    "df2['Solver'] = 'Dist.Greedy'\n",
    "df2['ar'] = df2['ar_db']\n",
    "df_cds = pd.concat([df1,df2], ignore_index=True)\n",
    "df_cds['Graph']=df_cds['type']\n",
    "df_cds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_patch(legend):\n",
    "    from matplotlib.patches import Patch\n",
    "    ax = legend.axes\n",
    "\n",
    "    handles, labels = ax.get_legend_handles_labels()\n",
    "    handles.append(Patch(facecolor='white', edgecolor='white', alpha=0.1))\n",
    "    labels.append(\" \")\n",
    "\n",
    "    legend._legend_box = None\n",
    "    legend._init_legend_box(handles, labels)\n",
    "    legend._set_loc(legend._loc)\n",
    "    legend.set_title(legend.get_title().get_text())\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,1,figsize=(8, 6))\n",
    "\n",
    "sns.lineplot(data=df_cds, x='size', y='ar', hue='Graph', style='Solver', \n",
    "             hue_order=['BA','ER','GRP','WS'],\n",
    "             style_order=['GCNN(5)+Dist.Greedy','Dist.Greedy'],\n",
    "             markers=True, ci=None, markersize=10, \n",
    "             ax=ax, legend='full')\n",
    "ax.grid(True)\n",
    "ax.set_ylim([0.9, 4.6])\n",
    "ax.set_xlim([90, 360])\n",
    "lgd = ax.legend(fontsize=12, framealpha=0.3, ncol=2, loc='lower left')\n",
    "add_patch(lgd)\n",
    "ax.set_xlabel(\"Graph Size\", fontsize=14)\n",
    "ax.set_ylabel(\"Approx. Ratio w.r.t. Central Greedy\", fontsize=14)\n",
    "fig.set_size_inches(6, 4)\n",
    "fname='./output/mwcds_ap_to_cgreedy_grp.pdf'\n",
    "# pp = PdfPages(fname)\n",
    "plt.tight_layout(pad=0.1)\n",
    "fig.savefig(fname,\n",
    "            dpi=300, #facecolor='w', edgecolor='w',\n",
    "            orientation='portrait',\n",
    "            format='pdf')\n",
    "# fig.savefig(fname.replace('.pdf','.png'),\n",
    "#             dpi=300, #facecolor='w', edgecolor='w',\n",
    "#             orientation='portrait',\n",
    "#             format='png')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Node Weighted Steiner Tree (NWST) results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>steiner_0</th>\n",
       "      <th>steiner_i</th>\n",
       "      <th>m</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BA</td>\n",
       "      <td>200</td>\n",
       "      <td>14</td>\n",
       "      <td>0.244142</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>80.241424</td>\n",
       "      <td>82.631852</td>\n",
       "      <td>0.971071</td>\n",
       "      <td>26.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BA</td>\n",
       "      <td>150</td>\n",
       "      <td>17</td>\n",
       "      <td>0.177841</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>68.674708</td>\n",
       "      <td>66.902516</td>\n",
       "      <td>1.026489</td>\n",
       "      <td>25.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BA</td>\n",
       "      <td>250</td>\n",
       "      <td>24</td>\n",
       "      <td>0.283682</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>34.688483</td>\n",
       "      <td>38.763846</td>\n",
       "      <td>0.894867</td>\n",
       "      <td>25.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BA</td>\n",
       "      <td>250</td>\n",
       "      <td>25</td>\n",
       "      <td>0.217516</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>55.979795</td>\n",
       "      <td>55.979795</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>29.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BA</td>\n",
       "      <td>250</td>\n",
       "      <td>13</td>\n",
       "      <td>0.173938</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>106.545693</td>\n",
       "      <td>106.026600</td>\n",
       "      <td>1.004896</td>\n",
       "      <td>29.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>2.0</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7992</th>\n",
       "      <td>WS</td>\n",
       "      <td>300</td>\n",
       "      <td>11</td>\n",
       "      <td>0.240148</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>39.394060</td>\n",
       "      <td>44.371841</td>\n",
       "      <td>0.887817</td>\n",
       "      <td>39.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7993</th>\n",
       "      <td>WS</td>\n",
       "      <td>250</td>\n",
       "      <td>13</td>\n",
       "      <td>0.176249</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>29.641437</td>\n",
       "      <td>29.497725</td>\n",
       "      <td>1.004872</td>\n",
       "      <td>26.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7994</th>\n",
       "      <td>WS</td>\n",
       "      <td>300</td>\n",
       "      <td>10</td>\n",
       "      <td>0.331071</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>33.071880</td>\n",
       "      <td>35.291784</td>\n",
       "      <td>0.937099</td>\n",
       "      <td>30.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7995</th>\n",
       "      <td>WS</td>\n",
       "      <td>100</td>\n",
       "      <td>16</td>\n",
       "      <td>0.312400</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.018567</td>\n",
       "      <td>8.018567</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7996</th>\n",
       "      <td>WS</td>\n",
       "      <td>250</td>\n",
       "      <td>27</td>\n",
       "      <td>0.338037</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.158072</td>\n",
       "      <td>8.126371</td>\n",
       "      <td>1.003901</td>\n",
       "      <td>12.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>7997 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     type  size   k         p  mwds  mwcds         gcn      greedy     ratio  \\\n",
       "0      BA   200  14  0.244142   NaN    NaN   80.241424   82.631852  0.971071   \n",
       "1      BA   150  17  0.177841   NaN    NaN   68.674708   66.902516  1.026489   \n",
       "2      BA   250  24  0.283682   NaN    NaN   34.688483   38.763846  0.894867   \n",
       "3      BA   250  25  0.217516   NaN    NaN   55.979795   55.979795  1.000000   \n",
       "4      BA   250  13  0.173938   NaN    NaN  106.545693  106.026600  1.004896   \n",
       "...   ...   ...  ..       ...   ...    ...         ...         ...       ...   \n",
       "7992   WS   300  11  0.240148   NaN    NaN   39.394060   44.371841  0.887817   \n",
       "7993   WS   250  13  0.176249   NaN    NaN   29.641437   29.497725  1.004872   \n",
       "7994   WS   300  10  0.331071   NaN    NaN   33.071880   35.291784  0.937099   \n",
       "7995   WS   100  16  0.312400   NaN    NaN    8.018567    8.018567  1.000000   \n",
       "7996   WS   250  27  0.338037   NaN    NaN    8.158072    8.126371  1.003901   \n",
       "\n",
       "      steiner_0  steiner_i    m  \n",
       "0          26.0       25.0  3.0  \n",
       "1          25.0       25.0  3.0  \n",
       "2          25.0       22.0  7.0  \n",
       "3          29.0       29.0  5.0  \n",
       "4          29.0       29.0  2.0  \n",
       "...         ...        ...  ...  \n",
       "7992       39.0       36.0  3.0  \n",
       "7993       26.0       26.0  2.0  \n",
       "7994       30.0       27.0  3.0  \n",
       "7995        8.0        8.0  5.0  \n",
       "7996       12.0       12.0  9.0  \n",
       "\n",
       "[7997 rows x 12 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "steiner_csv_dict = {\n",
    "    \"BA\":'./output/result_GCNSteinerGRP_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_BA_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"GRP\":'./output/result_GCNSteinerGRP_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_GRP_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"ER\":'./output/result_GCNSteinerGRP_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_ER_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"WS\":'./output/result_GCNSteinerGRP_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_WS_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "                   }\n",
    "\n",
    "df_s = pd.DataFrame([])\n",
    "for key in steiner_csv_dict:\n",
    "    steiner_csv = steiner_csv_dict[key]\n",
    "    try:\n",
    "        df_tmp = pd.read_csv(steiner_csv, index_col=0)\n",
    "        df_tmp['m'] = df_tmp['p']*df_tmp['k']\n",
    "        df_tmp['m'] = df_tmp['m'].apply(np.round)\n",
    "        df_tmp['type']=key\n",
    "        df_s = pd.concat([df_s, df_tmp], axis=0, ignore_index=True)\n",
    "    except:\n",
    "        pass\n",
    "df_s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbwAAAEsCAYAAAC41gacAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAABQDUlEQVR4nO3dd3zV9b348df7jOSckz0gCQQJkMiwqAwXKgJu617VOmutrWDt1Xrr+mm92q1dVq1VL8VFtY6q7bWoKME9AAcKAmGWFSALQvbJ5/fH93tOzklOck7IORnk/Xw8zuOc852ffDic9/lsMcaglFJK7e8cfZ0ApZRSqjdowFNKKTUoaMBTSik1KGjAU0opNShowFNKKTUoaMBTSik1KLj6OgGJkpuba4qKinp8nb1795KSktLzBKkONG8TR/M2cTRvEyceebt06dJdxpghkfbttwGvqKiIJUuW9Pg6paWlzJgxo+cJUh1o3iaO5m3iaN4mTjzyVkQ2drZPqzSVUkoNChrwlFJKDQoa8JRSSg0KGvCUUkoNChrwlFJKDQoa8JRSSg0KvRbwRGSuiOwQkS872S8icr+IlInIFyIyOWTfFSKyxn5c0VtpVkoptf/ozRLePOCULvafCpTYj2uAPwOISDbwU+AI4HDgpyKSldCUKqWU2u/0WsAzxrwNVHZxyFnAE8byIZApIgXAycAbxphKY0wV8AZdB06llFKqg/4008pw4D8h7zfb2zrb3oGIXINVOiQvL4/S0tIeJ6q2tjYu11Edad4mjuZt4mjeJk6i87Y/BbweM8Y8AjwCMHXqVBOP6X90GqHE0bxNHM3bxNG8TZxE521/6qW5BRgR8r7Q3tbZdqWUUipm/SngvQJcbvfWPBKoMcZsA14DThKRLLuzykn2NqWUUipmvValKSJ/A2YAuSKyGavnpRvAGPMw8CpwGlAG1AHfsfdVisg9wCf2pe42xnTV+UUppZTqoNcCnjHm4ij7DTCnk31zgbmJSJdSSqnBoT9VaSqllFIJowFPKaXUoKABTyml1KCwX43DU0opNfC0VFbSWlfP5MIRNG3egsPnxZWdHff7dBnwROQt4FxjTHXc76yUUkoBrXX1rD3hhOD7MQsXQvzjXdQS3gwgKf63VYNZb/2aU0qpUFqlqXpdb/2aU0r1X4EfvgAt27aF7wt5H88fxLEEvEIR8XR1gDFmU1xSo5RSalBo/8M31MbLLgu+jucP4lgC3idd7BPAAM74JEftr/ri15xSqu+0NjXRsmMnLTt20LKj3H7eQfOOHbSU72DIj67v9TTFEvBOBSoSnZD+pmXXDlobmsPbmbwO8Gb2ddL6JdPaiqmvx793L61799K6t47WurbXyQeWsP7MsyKeG/Zr7rUFtHq9OLze3kq6UqobjN9PS0WFHcxCAll5eUiA24G/qqrDueJ24xo6FNfQoYjL3etpjyXgfWaM2ZHwlPQnNVtorapk7RkXBjeNeWMBfPQQHHcz+AZ+CcT4/bTW1VmPYJDa2+59XeTtkV7X14Mxnd7vgCcejyldzeXlrD35FMTnw5WdjTMnG1d2jvWcZb/PycGZHfKclYW4e/8/j1L7E2MMrTU1NJfvCAatlh3lVoksEMjKy2nZtQtaW8NPdjhw5eTgysvDPXw43kmH4ho6FHdeXjDAuYYOxZmZiYgAVq3PmIULrdfbtoX98B355JO4CgqsS/vi9+NXO620V7MFHj8dZj0Uvr12B3z0MIyaDiUng7N3s860tEQMOP69ezF1dW0lq4iBqa7DdlNfH/O9xefDkeLD4fPhSEnB6UvBOSSXpKKROFJSgtu7ei2u2PLLmZnJkBtvxF9RQUtlJf7KSpq3b6fhq69oqayElpbI52Vk4MzJwZmd1RYgQ55dOdk4c3JwZWfjSE9HHDrngho8WvfuDQ9cdsksUL0YCHCmqanDuc6MDFx24EouKcGVNxR3SBBzDc3DlZMd8//xAFd2dqdtc66CApIKI67z3SPRUrgR8Mf9rv1Uy64dtOxtxJz1PC3bwwu1jbtb8H9nGWLAVVmBa0hel9dqbWoKLyXVhb7uolTVyetIH8SIRCIGHnd+fszBKfy1Ny7BoTu/5nKv+V7EaxhjaN29m5aKSvyVFW3PlZX4Kyrt5woa167F//HH+KurIyfG5cKZldkWCDt7zsmxSo8+X/BXqRq8+uNwGtPURMvOneGBa2egirEtkLXW1nY4V3y+YODyHnpoeCALlMyGDMGRnJyw9Fc1VFHXUke2Cf8h6zctbKndgs/lI8uTFbf7iemiGqrTk0SmA6nA+/11UPrUqVPNkiVLunVO0+YtnfYaClX03N+peuJxWhsaw9qr/Hv3YvbW4a+rg+bm2G7qdLYFlxQfDp/9HBJ4nCkpiM+HM4ZAJV5vv/9ybp/PYxYuTMivOdPSgr+qKhgIWyqrwgNlRWVbSbKigta6uojXEY/Hrl7NaXsODYxZ2eEBMqnvhq4GvpQbGhrweDz94kt5f9Fbn1uwmhz8lZUdSmDNwTYzq6Tmr4ywUprbjXvIkPDANXSIFcxCqhidqakJSXt3bKndwikvnML/Hv47hjoygtt3tNbw3Y9vZMF5Cxie2r08FpGlxpipkfZFm2nlOiDDGPPzkG3/wurIIsA2ETnBGLOyWyka4Ex9HXVvvowj2YXD58GRkoorIxPHAQfgyBqKIzMPR2pqx4CU0jFQSVJSvw9QA5W4XLiGDME1ZEhMx7c2NOCvrOxYggx5bt65g4ZVq/BXVGA6+VHjSE9vFyDt6tXs7A4lSGdGRlyrV3WMY/8WqKVo38EjrK0s0E7mb1e5JoIzNwf3kKG4CwrwHnKIFchC28ny8uL+meoN3/34xl65T7QqzSuAPwbeiMg5wMnA5cBK4AHgDuDbiUpgf7QnK4mnb59IVuNeMuqrydq9hcym1WS2tpLpbyWzMYm0lFFIcglklUBOCeTmQk4xeNL7Ovl9zuHzMmbhwrBSSH/g8HhwDBuGe9iwqMcaY2itrQ2WEFsqKuxq1ZDnyiqaNqynZelSq8dapNoUh8MKhBEDY/sOOjk4UrR6tTd1ZziNw+NpaxcLBC67ijG0pGYaGzvcx5GREaxOTB4zBlee9TqsijEnp9vtZH2pvqWemsaatkeT9VzdWM3uxt3UNNVwfsn5vZqmaLk3Bvg05P1pwCvGmKcBROQ2ILbudwOAw+el6I0FlNeV495Zw+6r28aJpD92P81DMhjizWVt41beaqmgpqUGv8sP2SlASti1XOwhfe8ysmo+ImONnyx/qxUQnR6yPLlkpBWQlX4AGdnFZA05iMwhE0jzZuGQgfXLbF8EGqvfLy1lxowZfZ2cfSIiONPScKalkVRUFPV44/fjr662AmNogKyqDAuU9V8ux19RGbHNBUCSk8N7rGbnkPXti61qVIcTf8WusOObt2zGtFglUXG7resGAq8xwdfBpo1ATA4GZxP+vqtzunNsF9fvVloC144lLZ0d28X1k0tKWHfGmUQS2g59wJNPsOmyyzscI15vWzvZwQfbVYrtqheHDMHh6XJujz4VGrh2N+2murG600BW01gTDGaN/o6BPSDJkURmciZnF5/de38I0QNeMrA35P1RwKMh79cBQ+OdqL7i8joob9nDGW99l38e+b9h+5qHZHDGh99lwUnzODhnPItLFmOMYU/zHmoaaqhqrKK6sZrqxmqqGqqoabS21dRXUbV3GxvrK/i8aTfV/gZaqIa91bB3JWx7LXgPh4FMcZLh8pKVlEmGL5es1GFkpOaRlZxFZnImmcmZZHmyyEjOICs5i7SkNJwOHfff34nTaXXbzsmJ6fjWxsa26tWqTkqQFZU0rikj49xz2HThtyJeZ9MVVwZfH/DE42y6/Ip4/DmDRqzDaRw+H0NuvLFDFaMjNbXflMgbWhraglL7wNVUExbEQkthsQSu9OR0MpIzGJk+kozkDDKSMqxn+5GZnEl6Unrwvddl1epsqd3SW38+EFsvzcOADSIyFBgPvBuyPx+oTkzS+oA3E5r3dH1MyhBIsb60RIT0pHTSk9IZwYiYbmGMYW/zXisY1myiaudKqqvWUl2zieq926mur6Bq705qHLvYXL2OrxwOqpxOmjv5TyNI8AOVmZxJpscOisl2UAwJjoF9GUkZfRIkAz2yAkZ8Y0TYBz7ePbIGMkdyMo6CAtx279WuNG3eHNM1XdnZDL/fbqGwP08iEnwdfEbCnsKODbzvcKy0e7uPx3Z6ThfHBp+7cf0O53RybKzDaTIyO+1dHG8NLQ1tQalpd1jpqqbJKmGFvq9pqOlR4EpPTre+NyIEskDgGiii/Ws+DjwoIhOxVk5YaYxZGrJ/GrA8QWnrG3Zj747WGoa+3Pbrbkdrjb2/Z4FCREhNSiU1KZURaSOg8OiOB/lboHojVJTBrjWYXaupr1hDVdVaqhsqqXY6qXI4qHG6qErJolqgmiaq/RVsq9vFytZGqhpraGqNPJRBENKS0sKDYRcBM/Bhdzl61n5Q11LHKS+c0un+BectIAsNeN0XWwlCPF7STzopwWnZv7RUVkZt5sjz5e1TO3QgcEUqXUUMXPb+7gSuEakjmJg7MWrg8jg9/aYkmkjRvsHuxWqcOgPYDny/3f6jgWcTkK4+43P5WHDeguD7QMeK4QxnwYgF+Fy+xCfC6YKcMdbjwJMRwGc/hjfU2IGwDCrWwK411vtNn0NL22Byk5xOfc5oarJGUpVRQHVqLtXedKrdHqpb9oZVu+6o28GqqlXUNNZQ39L5gPS0pLSIwTA0ULavdnU7dAaURAt0AoLem7FisHBlZ1OeVN9lM8e/z/03bkczNVWrO7RtBasGQwJZoJ2rwd/Q6X0HS+Dq7Ps2dH88dRnwjDGtwJ32I9L+C+Kamn4gy5MVVsooXdLPOlZ4MmD4FOsRqrUVdm8OBkCpKMO3aw2+zcso+Oo/4cdmjLB6jOaWQM5hkFts9SRNH05Da1PEtsjqxmqqG6qD+yoaKlhbvZaqxqoug2SqOzUYBG894tYu/7R9GROq+mbGiv1Zs7+ZioYKdtXvoqK+gmGpXffa3bZ3G1e9dlXEfW6Huy04RQhcgeaIgR649lVvf99GG4f3OPAm8JYxJraGAtU3HA7IPMB6FB8fvq+pDirXtpUGd62xSoef/Q2aQtosXV48OcXk5xaTn1NiB8RiKJjW5XCKRn9jWDBsHxwDAbPVtHZ6DbC+OM575TxyvDnkenPJ8eSQ483p9Nnn7oXSttov+Fv9VDVWUVFvBzI7oAUeFfUV1r6GXdQ01oSdO/fkuUDnzRzp7lTuPOrOQR24BopYhiVcBLhEZD3wFrAIKwCWJzpxKk6SfJA/0XqEMgZqy9sCYKCadOtnsOJlCA1Qqfl2qdAuDQaCYeZIkp3J5KXkkZfS9XRr0XpkpbtTObfkXCoaKqisr2R9zXqWlC+hurE64vFel5dsT3Z4cOwkQKa4UwbFF09/HeOYCMYYdjftDgax9oEsNLhVNlRG/MHlc/msz483h9GZoznMcxi53tzgtlxvLukua8hRZ4OjUx1uLhhzTq/Pr6u6L1qV5jEi4gOOweq0MhO4EnCKyNe0Bb8XE5xOlQgikJZvPUYdG76vpREq14e3E+5aYwXC+pBlP5xJkDWqLQDmlrQFxParSrR2PS1rqsPNzVN+3OGLo7m1maoG69d5RUNFxOdNezbx2c7PqGqowtCxajTZmUyOx/oCy/Zmd1569OaQ5k4bsMFxfxjjWNdcFyxtdQhegeDWYL1ubu04243b4Q4GrYKUAr6R+43g+2Ag8+TGVktQX82WPVEqtxpqIKkm2Htb9V9Rf5IYY+qA1+0HIpKKFQBPAr4LXIsuALv/cSXD0HHWo729FSGB0C4Z7loNq1+D0C8gb3ZbADzyWog2pr6TLw63w81Q31CG+qIP+WxpbbHaGOsjB8Zd9bvYWruV5TuXU9VYFfFXf5IjiWxvdvBLsatq1fSk9AEbHHtTk7+JyobKsGrEYFViu+rFSG3CDnGEleZHZ44OD2L2j5m4/JsYA811UF8NjbshOa3r40OGKqn+LeYyuIg4gClYpbwZwLHAHuDVhKRM9V8pOdbjgCPDtweGUwQDoV0yLHsDvv4nvqvfYMHpL0BDVXh1KYDDhS8lH1a+AsYPyRlWu2FyevhzUlpw6EgkLocr+CUYjb/VH+yAE7H0WF9BeV05KypWUNlQid90LKG6HK6O1aqdPGckZyRsJp2+GOPYvl0sLJCF/MDYVb+L3U27I14jIzmDXI/179W+JBb6gyMrOat7Y0eNgcY90FBt/ZCqr7Zed/lc0/Y65Ieb7+L5LDj1b+D2wN6d4fdJGYqvYj08ciKk5VlV/4Fak9S8kNf5kDoUnNpruS9F67QyGSu4zaItwC0GXgL+yxizOsHpUwNJ6HAK2o23a6ghq6KMrNZW64voqfPC91/xT/i/K2DLJ1FuItYv7vaBMOJzRufHOpw4Hc7gF2q04X+tppWaxppOS42BUsqqqlVUNlTS0tpx3T6XuMjyZEUtNeZ4cshMzuzWF3y8xjgG2sXCqg9DqhBDt3VWQg60i+V6cxmTOYbD8w/vWKXozSXbk02Ss4uVJVpbrRJWQzVUbY4cqDoNZjXWD6fOiMP6fHgyrQknPJmQOSL8vScDvJlkeXPIcqXCk+dA5brw65ScDMfeACUnWu3he7bB1k/twNi+al3AlwNpBSHBMc96HwiOgWdX4pbkGcyilfCWAJuA32IFuLLEJ0ntlwLDKWq2wOPXdNz/wUNwyd/B7YWG3fYX3W5orGn3PvS5xnqu3WGVJAPb/TGsHZiUGkPQbCtlOjzpZCWnkZWcTnFGCQyd0mknhdDOFKHBMbQKr6LeGtbRWTuUQxxkJWdF77Fql36iMa1+Nu3e1HnHDrvNLKZ2sdSO7WKhaQxrF2v120Gpqi0Y1ayGho9jKG3V0DFohGaSKzxA+bIhe1T4ts6ek1K7rCkIE1gUun2wA1jzGky5Ak7/Q/jnwd9iBb3a7bDHftSWh7zeDuUrrG2RArM3q4uAGBIwk7SncndEC3hzsUp49wIXiEgpVkeVD4wxnY+aVCqSrr44Vv8bNn1g/WJOy7Me+6q5ITwwRgyW7YJpXSVUbWh73xLDx9vtixgsxZNORnI6GZ4MRocGz6wRkB9yfHI6xulmT/OesDbHQACqbKgMbttQs4GKhoqIs2wIwl9P/muXSd1WV95hrFhYu5i3Y7tYMIglZZDe2oqEVvk1VFtBbM9/oH55hNKWfWxj5KrMIGdSeCBKzYPcsR0DlF3aCg9aKSFTkCVIfTW8f3/bZzZ7NE1nP0bSu/dan1mAt++DEUeGt+M5XZBeYD260uqHuorIATGwbcO71nOkH3LJ6eElw2BAbFelGq0dcpCI1kvzagARGYHVdjcTKwjmi8gnQClQaox5K8HpVAPdvn5x7Au3x3qk9mBe85am8FJkZ6XM9ttqNre9b468oGwocXlIT04n3ZPOqIglzOGQPR486ZikNPa6k6gQQwV+Klqb2NVSR0VdOZ4oX/zpRvj5lJ+Q60giByc5rZDV0owz0M5VXw27qqB+fcfSVvPeri4NLm94MEovhKEHRS9leTPB5Ul80OoJbyYcdzOMmg5v3wvfeprlZduZcvZD1g+0t++Fbz21759Zh9P6nEb7rBpj/8DYblWbBoJjoBp1Tzls/sR6jjQRhDslSjWqXWr0ZPbNv0d9tZXX0bb1UEydVowx/wGesB+ISBFWD80fAbfFeh01iCX6iyPeXEngyoWU6J1fOuVvtjtOdBU0I1TZ7t5mnde4G5ralgkSINV+jAy9jy+XLVdHXsImINWbzZlv/S5y6RqsL8TQQJRV1EmgyogQtPbz9iZftlXzYP8Y27NnTYdtCSdi3dOXDXkTOj/OGOvzFikg1m63nrd9AWveCPtsBbnsH4pdVaOm5Vs9sOO1yGzNFuvH8HE3R97WfnhTD3Snl2YhbaW8GUAR0IBVylMquv7wxdGbnO62L6l91ervuoQZCKYtnU8oDFjB98jZVttQpCCmvQe75nR1/HxG2tbXRKx/U28mDBnb9bGNtSEBMUKV6s5VsH6x3ZbajsNtB8IoPVNTcruecD+0mWPUdIYXjOmwjZKT4zaoP1ovzW/TNuB8NNAMfIi1isIi4ENjTAw9BJSyDZQvjv7C4bSClDdKx5Ro64q5fXB47yxfowaI5FTrkTOm6+Oa6ztpY7RLjVXrrVqa+sqO54rTGqcYKSCOmgEf/6Wt1uGlayk6/wl4/OK2bfFq5rBFC5vzgE+wVkRYBLynnVWU6n98OFlw4lyIMBwCwNfSaPUc1OmvVHe5vVbv1+xRXR/X0mj1mG7f6SZQpbp7C2xZ1jZkw5MBJ/8CznoQXrsNGmpwP3VW2/WyR8e9mSPapz/LGBOlxVop1afqq8l65/dkffSw9T5Sh6Bhk+GS57UkrRLHlWyNZcyMshi2v9kKeoGAmJwO58/tODb3wicgI74rfXTZ6hjvYCcip4jIKhEpE5FbIuwfKSJvisgXIlJqtxsG9v1GRL4SkZUicr/ofE5KWQIdgi6aD8MmwRX/YvlO4OyH2rb1pw5BanBzuiF9GAyfDGNPtTpIvfrfHY976+fWcKE4itaGt4cuR34CYIwxGdFuJCJO4EHgRGAz8ImIvGKMWRFy2H3AE8aYx0VkFvBL4DIRmYa12OzB9nHvAsehHWaUsgy2DkFq/xDr2Nze6LQCXNfFvoPs/bF27zocKDPGrAMQkWeAs4DQgDcBCKzBsQhrCjOwgq4HSMLqne0GdHkipUJphyA1kPTm2FxbtIHnj7ffJiIFwD3AFcACoEPVZCeGA6FLb28Gjmh3zOfAucAfgXOANBHJMcZ8ICKLgG1YAe8BY8zKGO+rlFKqv+mDsbndGYeXjhXcrgeWA8cbY96OW0osNwEPiMiVwNvAFsAvIsXAeCDQpveGiBxrjHmnXRqvAa4ByMvLo7S0tMcJqq2tjct1VEeat4mjeZs4mrfxlZM1jKKzH2fF15vZuXMnpXv2hG2rr18Tt3tFDXgi4gbmAP8PqACuNMY8vw/32gKEdt8ptLcFGWO2YpXwAuvunWeMqRaR72GN+au19/0bOAp4p935jwCPAEydOtXEYwHM0gG8kGZ/p3mbOJq3iaN5mxhHDB3ZIW+PGDqy8xP2QZe9NEXkEmAVcDNWwJuwj8EOrPF8JSIySkSSgIuAV9rdL9dedw/gVqx5O8FaseE4EXHZAfg4QKs0lVJKxSxaCe9JoB5r4PkE4PeRRgMYY66PdiNjTIuIXAe8hrVC+lxjzFcicjewxBjzCtasLr8UEYNVpTnHPv15rDX5lmN1YFlgjPln9D9PKaWUskQLeG9jBZiuhthHG7bQdqAxr9JuhXRjzJ0hr5/HCm7tz/MD34/1PkoppVR70XppzuildCillFIJ1e31HUTkaBHZz9cDUUoptb/ZlwWN/o01pk4ppZQaMPYl4OkclkoppQacOC1Zq5RSSvVv+xLwvo/OY6mUUmqA6fYU1MaY+YlIiFJKKZVI0ZYHmtvV/gBjzFXxSY5SSimVGNFKeEPavT8Za/7KusQkRymllEqMaAPPzwh9by8I+73AmnZKKaXUQKG9NJVSSg0K8Vk3XSmlVLc1NTXR0NBAa2trXyelX/D5fFRXV3e63+Fw4PF4SEpK2qfra8BTSqk+EAh2Pp8Pp9NJpJVoBhuv10tmZmbEfcYY/H4/dXVWF5J9CXrRemlObr8JOEhEwlJkjFnW7TsrpdQgFgh2LpeWO2IhIrhcLnw+H3V1dfEPeMASrOV/Qn96vNzuGIO1vp1SSqkYtba24nTqV2d3OZ3Ofa4CjhbwuloHTymlVA9oNWb39STPog1L2LjPV1ZKKaX6kS6HJYiIT0QeFJEtIrJDROaLSG5vJU4ppVTPlJeX8+1vf5vRo0czZcoUjjrqKP7xj3/E5dpFRUXs2rUrLtfqDdHG4f0PcCXwf8AzwInAnxOcJqWUUnFgjOHss89m+vTprFu3jqVLl/LMM8+wefPmsONaWlr6KIW9K1rAOxf4rjHmGmPM9cA3gbNFRFtalVKqn3vrrbdISkriBz/4QXDbyJEj+eEPf8i8efM488wzmTVrFscffzy1tbUcf/zxTJ48mYkTJ/Lyy1b/xA0bNjBu3DguueQSxo8fz/nnnx8cGgDwpz/9KXjO119/3et/Y3dEC3gjsObOBMAY8zHQAgxLZKKUUkr13FdffcXkye1Hl7VZtmwZzz//PIsXL8bj8fCPf/yDZcuWsWjRIn784x9jjAFg1apVzJ49m5UrV5Kens5DDz0UvEZubi7Lli3j2muv5b777kv439QT0QKeE2hqt60FHbCulFIDzpw5czjkkEM47LDDADjxxBPJzs4GrOrP2267jYMPPpgTTjiBLVu2UF5uLX06YsQIjj76aAAuvfRS3n333eA1zz33XACmTJnChg0bevGv6b5ogUuAp0SkMWSbB3hURIJlWmPMmYlInFJKqX130EEH8cILLwTfP/jgg+zatYupU6cCkJKSEtz39NNPs3PnTpYuXYrb7aaoqIiGhgag41CA0PfJycmANT6uv7cFRivhPQ5sBSpCHk8B/2m3TSmlVD8za9YsGhoa+POf2/oahra/haqpqWHo0KG43W4WLVrExo1to9I2bdrEBx98AMD8+fM55phjEpvwBIk2Du87vZUQpZRS8SUivPTSS9xwww385je/YciQIaSkpPDrX/+a+vr6sGMvueQSzjjjDCZOnMjUqVMZN25ccN/YsWN58MEHueqqq5gwYQLXXnttb/8pcbFPbXEicgCQCqw0gVZNpZRS/U5BQQHPPPNMxH1XXnll8HVubm6wFBdqw4YNuFwunnrqqYj7AqZOnUppaWlPk5tQ0Qaef0tErm237c/AemA5sFxEhicwfUoppVRcRGvD+yEQnKVTRE4Avg/cCVyA1YvzjoSlTimlVJ8qKiriyy+/7OtkxEW0Ks2xwIch788CXjfG/BxARBqABxKUNqWUUipuopXwUoGqkPfTgLdC3n8F5Mc7UUoppVS8RQt4m4GDAEQkHZgIvBeyPweoTUzSlFJKqfiJVqX5HHC/iPwSOAXYRngV51Sgf0+eppRSShE94N0DFAK/BbYDlxpj/CH7L8ZaSUEppZTq16INPK8HLu9i/8y4p0gppZRKgGhteEoppdR+QQOeUkqpmF155ZWISPCRm5vL6aefHnEtvOuvvx6n08mjjz7aByntSAOeUkoNUC99uoWjf/UWo275P47+1Vu89OmWXrnvCSecwLZt29i2bRuvv/469fX1nHPOOWHHNDY28vTTT3PLLbfw2GOP9Uq6ounVgCcip4jIKhEpE5FbIuwfKSJvisgXIlIqIoUh+w4QkddFZKWIrBCRot5Mu1JK9ScvfbqFW19czpbqegywpbqeW19c3itBLzk5mfz8fPLz85k8eTI33HADX3/9ddiE1C+++CJFRUXcfvvtrFixol/M1tJrC7mKiBN4EDgRa3zfJyLyijFmRchh9wFPGGMeF5FZwC+By+x9TwA/N8a8ISKphEx5ppRSA93//PMrVmzdHfPxn26qpskf/jVY3+znJ89/wd8+3hTTNSYMS+enZxzUrXS2t2fPHp599lkmTpyI1+sNbn/ssce49NJL8fl8nHfeeTz22GP84Q9/6NG9eqpHJTwRKbBXTojF4UCZMWadMaYJeAZrqrJQE2ibyWVRYL+ITABcxpg3AIwxtcaYyIs6KaXUINA+2EXbHk8LFiwgNTWV1NRU0tPTWbx4MfPnzw/uX79+Pe+88w4XX3wxAJdffjlPPfUUjY2NnV2yV/S0hPcWcCDWJNLRDMdaODZgM3BEu2M+B84F/gicA6SJSI59j2oReREYBSwEbmk3JhARuQa4BiAvLy8uS1XU1tb2+yUvBirN28TRvE2ceOWtz+cLKxGdWwTnFqXGfP5V5TXsrPN32D7E5+T/TYv1Oq0sX7485nsCVFVVMXnyZH76058CsHv3bp599llmzZrF008/TX5+Pn/60584/PDD2blzJzt37iQnJwe3283999/PKaec0um16+vrY0pPfX19pwvZdqWnAe9WIKOH1wh1E/CAiFwJvA1sAfxY6TwWmARsAp4FrgT+N/RkY8wjwCMAU6dONTNmzOhxgkpLS4nHdVRHmreJo3mbOPHK2+rqajIzM/f5/Ntbsrn1xeXUN7cFPa/bye1nTGTixMSt2paVlYXf7+eb3/xmcNuFF15IRkYGb7/9NnfddRevvvoqW7duZdKkScFjWltbef311/nv//7vTq+9fPlyJk6cGDUN+5p3PQp4xpiXunH4FmBEyPtCe1vo9bZilfCw2+nOM8ZUi8hm4DNjzDp730vAkbQLeEopNVicPckKave+toqt1fUMy/Ty3yePDW7vTSKCw+Ggrq6OBQsWUFFRwZIlS0hKSgoes2nTJk4//XQ2bNhAUVFRr6cRYgx4IvIWcK4xprrd9nTgJWPMrBgu8wlQIiKjsALdRcC3210vF6g0xrRilR7nhpybKSJDjDE7gVnAkljSrpRS+6uzJw3vkwDX2NjI9u3bAauK84EHHqC2tpYzzjiDP/7xj5x66qlMnjw57JxvfOMbjB07lrlz53L33Xf3epoh9k4rM4CkCNs9WFWNURljWoDrgNeAlcDfjTFficjdInJmyH1WichqIA/4uX2uH6u6800RWQ4I0D9GMiql1CCzcOFCCgoKKCgo4IgjjuCTTz7hueeeY/z48fzrX//i/PPPj3jeBRdcwF//+ldaW/umk32XJTwRCQ3RB4tIZch7J3Ay7aolu2KMeRV4td22O0NePw8838m5bwAHx3ovpZRS8Tdv3jzmzZvX6f7m5uZO99199919VrqD6FWaSwBjP16PsL8e+GG8E6WUUkrFW7SANwqr+nAd1ji6nSH7moAd7YcGKKWUUv1RtOWBNtovdc5NpZRSA1pvzrSilFJK9ZmeltzeAtbHIyFKKaVUIkUdhyciLqzpul6yB4aHivdMK0oppVRCRA14xpgWEbkX+L8I+15KRKKUUkqpeIu1SvNDYHLUo5RSSql+Kta5NB8FfisiI4GlwN7QncaYZfFOmFJKKRVPsQa8wEJHv4uwzxDb8kBKKaVUn4m1SnNUF4/RiUmaUkqp/ubKK69ERDo8jjzySACKioqC27xeL+PGjePee+/FGNPHKY+xhBcyAF0ppVR/sf5t+PdP4LKXYdeqttdpeQm97QknnMCTTz4Zti10KaA777yTa6+9loaGBhYuXMi1115Leno63//+9xOarmh04LlSSg1E69+G+RfCzjXwwtVtrxf/OuG3Tk5OJj8/P+yRnZ0d3J+WlkZ+fj5FRUVcffXVHHzwwbz+eqTpmHtXT1c8fws4EG3DU0qpnvn3LbB9eezHb10GzfXW643vgrGX3Fn2OOxcFds18ifCqb/qXjq7wRjD4sWLWblyJSUlJQm7T6x6OtPKrcBV8UiIUkqpbhj6DfBkgDjagp04ILs44bdesGABqampYY+bb745uP/2228nNTWV5ORkZs6ciTGG66+/PuHpiibWFc8fARYBi0NnW9GB50opFSfdLWkFqjRNyGKqDhcUHQ2nR+pQHz/Tp0/nkUceCduWmZkZfH3jjTfy3e9+l507d3L77bdz0kknMW3atISmKRaxVmn6gF8Dw0VkLVAaeESYbkwppVSi/fsn0GIvturygvGDvwlWvJTwgOfz+Sgu7rwkmZOTQ3FxMcXFxbzwwguUlJRwxBFHMHPmzISmK5qYqjSNMZcaYw4AxmIFPi/wS+A/IhJjZbFSSqm4uexlmHIF+HLgnIdh0mXW6wvm9XXKwmRlZXHddddxww039PnQhO624a0DvgRWAKsAP5DU5RlKKaXiLy3PKsn9ZB0cdHbb61HTE37rxsZGtm/fHvbYuXNnp8fPnj2bVatW8dxzzyU8bV2JKeCJyE9E5FWgGvgbVs/Mp4ESY8yoxCVPKaVUf7Nw4UIKCgrCHpMmTer0+KFDh3LZZZdx11130dra2ulxiRZrG96vgJ3APcA8Y0znoVwppdR+a968ecybN6/T/Rs2bIi4vX0nl74Qa5XmicAjwJnAJhFZLiJ/EpFzRSQncclTSiml4iPWqcXeBN4EEBEvMA24BKt60wG4E5VApZRSKh5inmlFRIYCM4EZ9vOBwHZgcUJSppRSSsVRrAPPV2IFuHKsAPd7rDF4OiRBKaXUgBBrCe8PaIBTSik1gMXahveXRCdEKaWUSqSeTh6tlFJKDQga8JRSSg0KGvA6Ud/sp3x3AyPGHkz57gbqm/19nSSllFI90NMFYPdL22saeHBRGc8v3Ux9sx+v28n5UwqZM7OY/AxPXydPKaXUPoi5hCciySJylYjcJyL3isiVIpKcyMT1he01DZz70Hs8+eHGYKmuvtnPkx9u5NyH3mN7TX0fp1AppfpWeXk5N9xwAyUlJXg8HoYOHcq0adP405/+RG1tLQBFRUWICCKC1+tl3Lhx3HvvvWErJmzYsCF4jIhw9NFHM336dBYvTszw7lgnj54ArAZ+BxwBHIk1VGG1iIxPSMr6QH2znwcXlbG1piHi/q01DTxUupYGrd5USg1SGzZsYPLkySxYsIB77rmHZcuW8dFHH3Hbbbfx5ptv8sorrwSPvfPOO9m2bRsrV67kpptu4rbbbos4p+aCBQvYtm0bf/3rX8nIyOC0005j/fr1cU97rFWafwQ+Ay4zxuwGEJF04CmswHdy3FPWB3bXN/P80s1dHvPcks3MmVmMx+3spVQppVRkM56dQUVDRfB9jieH0m+VJvSe1157LQ6HgyVLlpCSkhLcPmrUKE4//fSwElxaWhr5+fkAXH311fz5z3/m9ddf5/vf/37YNXNycsjPz+fAAw/k4YcfprCwMOJxPRVrwDsaOCwQ7ACMMbtF5Hbgw7imqA+1+Fujdk6pb/azvaaBee9vYGxeGiV5qYwZkqoBUCnVI7/++Nd8Xfl1t84JDXaB999Z8J2Yzx+XPY6bD7859vtVVPDaa6/xi1/8IizYhRKRDtuMMSxevJiVK1dSUlLS5T28Xi8Azc3NMacrVrEGvAYgM8L2DHvffsHldOB1O7sMel63k2Z/K4+9s45mv/VLxiFQlJPCgXlpHJiXyoH5aYzNS6MoNwW3UzvCKqX2D2VlZRhjGDt2bNj2wsJCqqurAbj00kt5+OGHAbj99tu56667aGpqorm5GY/Hw/XXX9/p9evq6rjttttwOp0cd9xxcU9/rAHvn8CjIvI92kp0RwF/AV7p9KwBJt3r5vwphTz54cZOj7lgaiHfGJ7BirtPYcOuvawq38Pq8lpWb9/D6h17eH3FdlrtEr3bKYzOTaUkL9UuDaYxNj+NA7J9OB0dfwUppQav7pS0AiY+PrHDtr+e8td4JKdb3nnnHfx+P9dccw0NDW1loBtvvJHvfve77Ny5k9tvv52TTjqJadOmdTh/+vTpOBwO6urqKCgoYN68eUyc2PFv66lYA96PgMeBd4BA8ceBFez+K9abicgpWO2BTuAxY8yv2u0fCcwFhgCVwKXGmM0h+9OBFcBLxpjrYr1vrLxuJ3NmFvPmyvKIHVeGZXiYPaOt/a7EDmKhGpr9rN1Zy5ryWlaV72FN+R4+31zNv77YFjwm2eWgeGhoEEylZGgawzO9ODQQKqVilOPJ6dCGl0jFxcWICF9/HV71OmrUKAB8Pl94+nJyKC4upri4mBdeeIGSkhKOOOIIZs6cGXbc/PnzmThxIlu3buXYY49NWPpjnUuzGjhLREqAcfbmlcaYslhvJCJO4EGsxWQ3A5+IyCvGmBUhh90HPGGMeVxEZgG/BC4L2X8P8Has99wX+RkeXpw9jYdK1/LckrZxeBdMLWT2jOjj8DxuJwcNy+CgYRlh2/c2tlC2oy0Iriqv5YN1Fbz46ZbgMSlJTkoC1aJ5aRxolwiHpiVHrBdXSg1uie6g0l5OTg4nnXQSDzzwAD/84Q9JTU2N+dysrCyuu+46brjhBj799NOw77TCwkLGjBlDXV1dIpId1K2B58aYNcCafbzX4UCZMWYdgIg8A5yFVWILmADcaL9eBLwU2CEiU4A8YAEwdR/TEJP8DC+3nTaeOTOLqd1bR2qKjwyvu0cdU1KSXRwyIpNDRmSGba+pb2ZNoFq0fA+ry/fw1tc7+PuStt6iGV53WBAMtBXmpO53wyCVUv3cQw89xNFHH82UKVO46667OOSQQ3C5XCxdupTPP/+ck046qdNzZ8+eza9//Wuee+45Lrzwwl5MtaXTgCci9wO3GmP22q87ZYzpvBWyzXDgPyHvN2ON6Qv1OXAuVrXnOUCaiOQAVcBvgUuBE7pI8zXANQB5eXmUlpbGkKyu1dbWdutXzL4aBgxLhxnpQImb3U0utuxpZUut/aip4R+bq6hraTsnPQmGpzoYnuqgMM16HpbqIMU9MEqDtbW1cfk3Uh1p3iZOvPLW5/MFeyQONPPnz+exxx7j5ptvZvv27bhcLkaPHs25557LxRdfzPLly2lubmbbtm0sX7487NxvfvOb3HrrrYwdO5Zt26ymnrKyMpKTk6mvr+9wfCT19fX7VBrsqoQ3EXCHvO4NNwEPiMiVWFWXW7DaDGcDrxpjNndVtWeMeQR4BGDq1KlmxowZPU5QaWkp8bhOPBhj2LGnkVXb9wRLg6vKa/lg+x72bmoKHpef7rF7iqZabYT28AlfUv+aSa4/5e3+RvM2ceKVt9XV1WRmZvb4On2lfTtce1u2bIm4/fnnnw++PuSQQ8LG7S1fvjymzir7mnedfgMaY2ZGet0DW4ARIe8L7W2h99yKVcJDRFKB84wx1SJyFHCsiMwGUoEkEak1xtwSh3QNGCJCXrqHvHQP0w8cEtze2mrYUl3Pmh17WLW91m4j3MPj6ypoamkNHjci28uBQ9OCwyZ0DKFSajCJ6Se/iNwJ3GeMqWu33Qv8tzHm7hgu8wlQIiKjsALdRcC3210vF6g0xrQCt2L12MQYc0nIMVcCUwdbsOuKwyGMyPYxItvHrHF5we3+VsOmyjpWbd8TDIKry/ewePVOWlp1DKFSanCJtY7rp8DDQPtKU5+9L2rAM8a0iMh1wGtYwxLmGmO+EpG7gSXGmFeAGcAvRcRgVWnOiTF9KgKnQxiVm8Ko3BRO+UZ+cHtTSysbKvZa1aLb2wJhb40hrG/2s7u+Obj0UrrXjVdLmXGheatU52INeAKYCNsnYY2Xi4kx5lXg1Xbb7gx5/TzwfPvz2h0/D5gX6z1VR0kuR7C3Jwe3bQ+MIVwdMpj+s//EdwyhLr2UOJq3SnWty4AnInuwAp0B1tklrwAn4MEq+an9QFdjCNfsqA0rEb6/tvtjCANLL4UO6g8svfTmynJenD2N/IyB2Wutr2neKhVdtBLedVilu7nA7UBNyL4mYIMx5oMEpU31EynJLg4dkcmh7ccQ1jVbHWXsQLi6vJY3V0YeQ3jJESP5ZENl1KWXbjttvHai6aZYl7XSvFWDXZcBzxjzOICIrAfeN8bEf/pqNWBl+NxMLcpmalF22PZdtY2sLt8TnF5t9fY9FGR4eHFZ5G7KAc8t2cwZhwzjlhe+CNse6ywzsRwV64Q1EtPVYr9ebNfat7/znrO/octaKRWDWKcWCy4/KyL5QFK7/ZvinC41gOWmJpObmsy0MbnBbVuq6mJaeinZ5WBcQXrbxkgtxxGYGA40sV6rL+4Z22ERr+dySEx5W1HbxIptu5l8QBYZXneXxyu1P4p1WEI68CfgQtoFO5v+bFRdinXppbx0Dw9+e3IvpmzgK9/dEFPeVtc38Z2/foIIHDg0jSlFWUwdmcXUkdmMyPbqfK1qvxdrL83fAocAZwMvAldhTRX2I+DHCUmZ2q/EuvSSljy6L9a8nTgsg/lXH8GSjVUs2VjFPz/byvyPrMqZIWnJTB2ZxZSRWUwtyuagYek6DlPtd2INeKcCFxtj3hERP7DUGPOsiGwDvk+UoQRKdXfpJRW7WPM2zetmWnEu04qtqmZ/q2F1+R6WbKxi6YZKlmys4t9fbgfA43Zw6IhMpo7MZkpRllaDKgD+8pe/cOONN1JVVUVSklXZ19TURGZmJqNHj+bLL78MHltWVkZJSQkLFy5k9OjR3HHHHZSWlrJz505ycnKYNGkSP/vZz5g0aVKvpT/WgJcJBH4+1gA5QBnwAfBY/JOl9kc9XXpJdW5f8tbpEMYXpDO+IJ3LjhwJWNWjSzZUsWRjJUs2VPHnxWvxLzJaDdqPrT7mWPy7duHMzeXAd99J6L1mzpxJXV0dH3/8MccccwwAH330ERkZGaxZs4adO3cyZIg17eGiRYtITk7m8MMPZ9KkSYwZM4a///3vDB8+nK1bt/L6669TWRnzMO64iDXgrQVGA5uAlcBFIvIx1ryXvZtiNaAlYuklZYlH3uale/jmwQV88+ACwBqD+fl/qrUatB/z79oV9pxIBx54IMOGDWPRokXBgLdo0SKOP/541q9fT2lpKRdccEFw+1FHHcXatWtZu3YtCxYsoLi4GICRI0dy1FFHJTy97cUa8OZhzclRCvwK+BfWGD0HVjueUjHzuJ143E5WLvtQZ/SPs3jnbUqyS6tBe8n2X/yCxpVfRz/QVvfpp9DSYo2NMQZEWDluPLhc+GKsJkweP478227rVjpnzpzJokWLuOOOOwArsF1yySWMHDmSRYsWBQNeaWkpP/jBDxgyZAgOh4MXXniBH//4x7hcfbdqS6zDEn4f8votERmHtQjrGmNM9MWLlFL7Ba0G7Uda7MUxA2NVAs8tLZGPj5OZM2dy3XXX0djYiDGGDz74gEcffZQDDjiAH/3IKv98/fXXbNu2jVmzZjF8+HDuv/9+fvKTn3DPPfcwZcoUpk+fzkUXXcRBBx2U0LS2t0+h1h53twlARC4yxjwT11QppQYMrQaNj+6WtAJtd5KUhGlqCj47c3MZ+eQTCUolzJo1i4aGBj744AOMMQwZMoTi4mIKCgpYu3Yt27dvZ9GiRfh8Po44wlrje86cOVx++eUsWrSIjz76iJdffplf/epXzJ07l8suuyxhaW0vasATERcwFmg2xqwO2X421ioJYwENeEopQKtBe0ugg8rKceMBME1NjP96ZcLvO2rUKEaOHElpaSnGGI477jgAUlJSmDJlCqWlpZSWlnLMMcfgdrf9W6alpXHmmWdy5pln8rOf/YyTTz6ZO+64o/8EPBGZgNVeN9J+/zLwA6wANxmrh+Y3E5xGpdQAFqkadHtNQ7AKdOlGrQbtCWdubrCXZm8JtOMZY7j88suD22fMmMFbb71FaWkpN954Y6fniwjjxo1j2bJlvZHcoGglvF8B64HrgUuAbwETgPnAWcaYPYlNnlJqf5Sf4eH0g4dx+sHDAK0G7YlED0WIZObMmcyfPx+AuXPnBrcfd9xxXHjhhezZs4eZM2cC8Nlnn/HTn/6Uyy67jAkTJpCUlMTixYuZO3cuF198ca+mO1rAOxw4zRizTETexQp49xljdOydUiputBp0YJk5cyZNTU0UFhYGhxoAHHPMMdTX15Oens6UKVMAKCwsZPTo0dx9991s2LCB1tZWDjjgAG666SZuueWWXk13tIA3FNgCYIypFpE6rJXIlVIqYbQatH8bMWIEJsJM5qmpqTQ3hy+qk5uby+9///sOx/aFaAHPAK0h71sBXSJIKdXr9qUadGpRNlNHZjFBq0EV0QOeEL7SeSrwRbuVzzHGpHc4UymlEqg71aBet5NDRmT0qBq0vtnP7vpmRow9mPLdDaR73Xh1hqABJVrA+06vpEIppXookdWg22saeHBRGc8vbZun9PwphcyZqXPADiQxrXiulFIDUTyqQbfXNHDuQ++FrURR3+znyQ838ubKcl6cPY38DG+f/H2qe/puUjOllOplXVWDLtlglQRDq0FvOvlA1u6sjbjsEsDWmgYeKl3LbaeN36cJ0I0x2rmmmyJ1lomVBjyl1KAVrRp00ogs7nttdZfXeG7JZubM7P5ajg6HA7/f36eTKQ9Efr8fh2PfOiBpTiulVIjQatAtVXXUN/u7PL6+2c/exhYaW/wku2IPeh6Ph7q6Onw+H06nU0t6URhj8Pv91NXV4fHsW7upBjyllOqEy+nA63Z2GfS8bifbdzdw2v3vcFhRNkcX53JMcS4TCtJxODoPYoEVw+vq6mhtbe30uMGkvr6e6urqTvc7HA48Hk8w77pLA55SSnUi3evm/CmFPPnhxk6PuWBKIWnJLi467ADeK9vFr/5trWmX5XMzbUwu04pzOKY4lwOyfR1KcUlJSfv85b0/qqurIzMzM2HX71HAE5GpgM8Yo7OvKKX2O163kzkzi3lzZXnEjivDMjzMtocmTCzMBKz1Ad9fu4t311TwXtku/m/5NgAKs7wcPSaXo0tymTYmh9zU5N78UxQ9L+E9CRwI6OhLpdR+KT/Dw4uzp/FQ6VqeW9I2Du+CqYXMntFxHF5euodzJhVyzqRCjDGs27WX98p28V7ZLl79chvPLvkPAOPy0zim2AqAhxdlk5KsFW6J1tMcvhzwxSMhSinVX+VneLnttPHMmVlM7d46UlN8ZHjdUXtmighjhqQyZkgqlx9VRIu/lS+37g4GwCc+2Mhj767H7RQmjciy2v9Kcji4MFOnQkuAHgU8Y8wn8UqIUkr1Zx63E4/bycplHzJjxox9uobLaa30cOiITObMLKa+yc+SjZW8V2ZVf/7hzdX8fiGkJDk5cnSOHQBzKRmaqr044yCmgCci5xljXuhk383GmF/HN1lKKbX/8yY5ObZkCMeWDAGgam8TH66r4F27BPjm1zsAaxaYo8dYAfDo4lyGZerMLvsi1hLeUyJyKnC9MaYOQEQKsdrwxgMa8JRSqoeyUpI4dWIBp04sAGBzVR3vl1kB8N2yXbz02VYARuem2MEvh6NG55Lh0/UAYxFrwDsCa5Xzz0Tk28AY4M/AR8AhCUqbUkoNaoVZPi48zMeFh1nrz60q38O7a6zS3wvLNvPkhxtxCEwcnhEs/U0ZmbVP05wNBjEFPGPMF/YQhIeAD7DWybvJGHN/IhOnlFLKIiKMy09nXH46Vx87mqaWVj7fXB0MgI+8vY6HSteS7HJwWFF2cPzfQcMycHYxAH4w6U6nlUOA44AyYARwuIikGWP2JCRlSimlOpVkB7bDirK54cQDqW1s4eP1Fby7poL31+7iNwtW8RtWkeF1c9ToHI4usWaAKcrpOAB+sIi108qdwO3Ag8AtwCjgaWC5iFxmjHkncUlUSikVTWqyi1nj8pg1Lg+AHXsa+GBtRbAEuOAraxWIYRmeYO/Po8bkMDRt8KznF2sJ7wfAGcaY1+33q0TkSOBnwEIgpikDROQU4I9YA9UfM8b8qt3+kcBcYAhQCVxqjNksIoditRmmA37g58aYZ2NMu1JKDTpD0zycdehwzjp0OMYYNlTU8W7ZLt4v28XrK8p5bulmAMbmpQXH/x0+KofU/XgAfKx/2cHGmF2hG4wxLcAtIvJqLBcQESdWCfFEYDPwiYi8YoxZEXLYfcATxpjHRWQW8EvgMqAOuNwYs0ZEhgFLReQ1Y0x1jOlXSqlBS0QYlZvCqNwULjtyJP5Ww1dba4Lj/576aCNz31uPyyEcOiIz2AHm0BGZJLn2nwHwsXZa2SUiycAlwASsTisrgPndmEfzcKDMGLMOQESeAc6yrxMwAbjRfr0IeMm+f3BBKmPMVhHZgVUKrI7x3koppWxOh3BwYSYHF2Zy7YwxNDT7WbqxKjgDzP1vreGPb67Bl+TkiFHZwQA4Ni+tyxUg+rtY2/AmAAuwqhSX25u/B9wlIqcYY1bGcJnhwH9C3m/GGu4Q6nPgXKxqz3OANBHJMcZUhKTlcCAJWBtL2pVSSnXN43YGgxpATV0zH6yrCAbARausr/jc1CSmjbHG/x1dnEth1sCaWVJiWS5dRN7Aqla8zBiz296WDjwFeIwxJ8VwjfOBU4wxV9vvLwOOMMZcF3LMMOABrE4xbwPnAd8IVF2KSAFQClxhjPkwwj2uAa4ByMvLm/LMM89E/duiqa2tJTU1tcfXUR1p3iaO5m3iDMa8rahvZUWFnxUVfr6qaGV3kxU38nzC+BwnB+U4GZ/tJDWpZ6W/eOTtzJkzlxpjpkbaF2vAqwMOM8Z81W77ROBDY0xKDNc4CrjLGHOy/f5WAGPMLzs5PhX42hhTaL9Pxwp2vzDGPB/tflOnTjVLliyJdlhUpaWl+zxvnuqa5m3iaN4mzmDPW2MMa3bUBnt/frS+ktrGFkTgoGHpwQVwDyvKjnkAfH2zn931zcGJudO9brz7OHheRDoNeLF2WmkAMiNsz7D3xeIToERERgFbgIuAb7dLaC5QaYxpBW7F6rGJiCQB/8Dq0BI12CmllEoMEeHAvDQOzEvjqmNG0exv5YvN1cH1/+a+u56/LF5HksvBlAOyOKbEqiqdODzyAPjtNQ08uKiM55e2Lb10/pRC5szsuPRST8Ua8P4JPCoi3wMCVYlHAX8BXonlAsaYFhG5DngNa1jCXGPMVyJyN7DEGPMKMAP4pYgYrCrNOfbpFwLTgRwRudLedqUx5rMY06+UUioB3E4HU0ZmM2VkNj86oYS9jS18vKGS99bs4r21Fdz72irufW0VaR4XR43O4ZiSXKaNyWXMkBTKdzdy7kPvhS2uW9/s58kPN/LmynJenD2N/Iz4TZQda8D7EfA48A7WODgAB1aw+69Yb2aMeRV4td22O0NePw90KMEZY57Cai9USinVj6Uku5g5digzxw4FYFdtI++vreD9sl28s8YaAwhw3cxiKvY2RlxJHmBrTQMPla7lttPGx21u0KgBT0QcwDCsIQn5WKsjAKw0xpTFJRVKKaX2S7mpyZx5yDDOPGQYxhg2VVoD4CcOz+Bbf+nQ9zDMc0s2M2dmce8FPKwxd58BE+wAp0FOKaVUt4kII3NSGJmTwpaqOuqb/V0eX9/sp8XfGrf7Rx1Cb6xunKuwBnorpZRSPeZyOqL2xPS6nbic8ZvpJdYr/QS4T0QOlcE6zbZSSqm4Sfe6OX9KYZfHXDC1kAxv/Ba3jTXg/R1rarClQIOI7A59xC01SimlBgWv28mcmcUM62TowbAMD7NnxK/9DmLvpXld9EOUUkqp2OVneHhx9jQeKl3Lc0vaxuFdMLWQ2TP6aByeMebxuN5VKaWUAvIzvNx22njmzCwOzrSS4XXHtWQXEPPCRyLiwZoZZYK9aQXwN2NMfdxTpZRSatDwuJ143E5WLvswodO2xdSGJyKTgXXAb7Ha8g7HWrtunb1PKaWU6tdi7bTyCPAuUGiMmW6MmQ6MwJr+65FEJU4ppZSKl1irNA/CWnF8b2CDMWZvYB7MhKRMKaWUiqNYS3hfY00v1l4BsDrCdqWUUqpfibWE9/+A++0SXWDysyPt7beISHbgQGNMZXyTqJRSSvVcd5YHApiPNbcmQGDGlZdD3huspX+UUkqpfiXWgDczoalQSimlEizWgeeLE50QpZRSKpG6M/C8ALiWtoHnK4E/G2O2JiJhSimlVDzFOvD8RGAt8C2gzn5cAJSJyEmJS55SSikVH7GW8O4HHgN+ZK+PB4CI/BH4I22roCullFL9Uqzj8IqAB0KDne1BYGRcU6SUUkolQKwBbwkwMcL2icCn8UuOUkoplRixVmk+BPxeREoIH3h+LdbA8+AE0saYZfFNolJKKdVzsQa8p+3nX3SxD3TguVJKqX4q1oA3KqGpUEoppRIs1oHnGzvbJyInGGMWxi9JSimlVPzFPPA8lIgMB74DXIXVS1OrMZVSSvVrsfbSREScInKuiPwfsAE4B3gYKE5Q2pRSSqm4iVrCE5GxwNXA5cBerBUTTgIuM8asSGzylFJKqfjosoQnIu9gDUPIAi40xow2xvy/XkmZUkopFUfRSnhHYc2m8ogx5qteSI9SSimVENHa8A7DCorvisinInKDiOT3QrqUUkqpuOoy4BljPjXGzAEKgN8BZwL/sc/7pohkJT6JSimlVM/F1EvTGNNgjHnSGDMTa2WEe4EbgO0i8u9EJlAppZSKh5iHJQQYY8qMMbcAI4ALgaa4p0oppZSKs30aeA5gjPEDL9sPpZRSql/rdglPKaWUGog04CmllBoUejXgicgpIrJKRMpE5JYI+0eKyJsi8oWIlIpIYci+K0Rkjf24ojfTrZRSauDrtYAnIk6sQeynAhOAi0VkQrvD7gOeMMYcDNwN/NI+Nxv4KXAEcDjwUx0SoZRSqjt6s4R3OFBmjFlnjGkCngHOanfMBOAt+/WikP0nA28YYyqNMVXAG8ApvZBmpZRS+4l97qW5D4ZjDVoP2IxVYgv1OXAu8Ees1RjSRCSnk3OHt7+BiFwDXAOQl5dHaWlpjxNdW1sbl+uojjRvE0fzNnE0bxMn0XnbmwEvFjcBD4jIlcDbwBbAH+vJxphHgEcApk6dambMmNHjBJWWlhKP66iONG8TR/M2cTRvEyfRedubAW8L1mD1gEJ7W5AxZitWCQ8RSQXOM8ZUi8gWYEa7c0sTmVillFL7l95sw/sEKBGRUSKSBFwEvBJ6gIjkikggTbcCc+3XrwEniUiW3VnlJHubUkopFZNeC3jGmBbgOqxAtRL4uzHmKxG5W0TOtA+bAawSkdVAHvBz+9xK4B6soPkJcLe9TSmllIpJr7bhGWNeBV5tt+3OkNfPA893cu5c2kp8SimlVLfoTCtKKaUGBQ14SimlBgUNeEoppQYFDXhKKaUGBQ14SimlBgUNeEoppQYFMcb0dRoSQkR2AhvjcKlcYFccrqM60rxNHM3bxNG8TZx45O1IY8yQSDv224AXLyKyxBgzta/TsT/SvE0czdvE0bxNnETnrVZpKqWUGhQ04CmllBoUNOBF90hfJ2A/pnmbOJq3iaN5mzgJzVttw1NKKTUoaAlPKaXUoDDoA56IzBWRHSLyZci2bBF5Q0TW2M9Z9nYRkftFpExEvhCRyX2X8v6tk3y9S0S2iMhn9uO0kH232vm6SkRO7ptUDwwiMkJEFonIChH5SkR+ZG/Xz20PdZG3+tntIRHxiMjHIvK5nbf/Y28fJSIf2Xn4rL1eKiKSbL8vs/cX9TgRxphB/QCmA5OBL0O2/Qa4xX59C/Br+/VpwL8BAY4EPurr9PfXRyf5ehdwU4RjJwCfA8nAKGAt4Ozrv6G/PoACYLL9Og1Ybeehfm4Tl7f62e153gqQar92Ax/Zn8e/AxfZ2x8GrrVfzwYetl9fBDzb0zQM+hKeMeZtoP1ismcBj9uvHwfODtn+hLF8CGSKSEGvJHSA6SRfO3MW8IwxptEYsx4oAw5PWOIGOGPMNmPMMvv1HqwFlYejn9se6yJvO6Of3RjZn79a+63bfhhgFm3roLb/3AY+z88Dx4uI9CQNgz7gdSLPGLPNfr0da/V1sD74/wk5bjNd/2dQHV1nV6vNDVS5ofm6z+xqnklYv5b1cxtH7fIW9LPbYyLiFJHPgB3AG1gl4mpjTIt9SGj+BfPW3l8D5PTk/hrwojBWeVq7ssbHn4ExwKHANuC3fZqaAU5EUoEXgP8yxuwO3aef256JkLf62Y0DY4zfGHMoUIhVEh7Xm/fXgBdZeaDKx37eYW/fAowIOa7Q3qZiYIwptz/wrcCjtFX9aL52k4i4sb6QnzbGvGhv1s9tHETKW/3sxpcxphpYBByFVcXusneF5l8wb+39GUBFT+6rAS+yV4Ar7NdXAC+HbL/c7vV2JFATUoWkomjXbnQOEOjB+Qpwkd0raxRQAnzc2+kbKOx2jP8FVhpjfheySz+3PdRZ3upnt+dEZIiIZNqvvcCJWG2ki4Dz7cPaf24Dn+fzgbfsmot919c9d/r6AfwNq4qiGav++LtY9cRvAmuAhUC2aetl9CBWvfNyYGpfp7+/PjrJ1yftfPvC/jAXhBx/u52vq4BT+zr9/fkBHINVXfkF8Jn9OE0/twnNW/3s9jxvDwY+tfPwS+BOe/torB8JZcBzQLK93WO/L7P3j+5pGnSmFaWUUoOCVmkqpZQaFDTgKaWUGhQ04CmllBoUNOAppZQaFDTgKaWUGhQ04Cm1nxCRUhF5IAHXvVJEaqMfqVT/pgFPqTgQkTwR+b29NE+DvTTS+yLyQ3uaqn5LRK4WkU9FpFZEauz5In8WcsizWGOllBrQXNEPUUp1xZ5k+D1gN3AH1sDaeuAg4Gqs6ZDmd3JukjGmqXdSGvH+VwH3AzdgDVp3A9/AmvIJAGNMPdbfo9SApgPPleohEfk3VpAYZ4zZG2G/GPs/mogY4DrgeOBkrEmJbwYewVomJR9rZppHgfuMNXcjIjIPyAU+BH4IpGDNQjHbDkiISCmwAqgGrgFagSeAnwSuEyFtLwG1xphLu/j7rgQeMMakhvwNHRhjxN4/HGty5cBiqO9jTcK8prN7KNUbtEpTqR4QkRysL/YHIwU7CK5cEOqnwKvARKwpvxxYE+VeCIzHmqrqNuA77c47DjgEK1ieB5wE/LrdMZcALcA0rMD6X8C3uvgTtgOHi0h3qiwLQh4jgKXAYgAR8WHNjdhgp/corCnmFtr7lOozWqWpVM8UY81VuSp0o4hsBjLtt08ZY34QsvtZY8xj7a5zZ8jrDSIyGbgYayLjAD/wHWMtovmliNwM/K+I3BoSbFcYYwLXWi0i38MKkH/rJP3/gxVE14pIGdbab68DfzPGNEc6wRizPeTvfMj+OwOluYuw8uM7IaXa72Ot3HA61urWSvUJLeEplRjHYq2d9jHWJLihlrQ/WER+ICJLRGSn3SPyBuCAdod9YdpWjAb4AEjCWqcteEy7c7YCQztLpLFW+D4Kq7T5B6xg9Rfg42glMhGZA3wbOMMYE1i2ZQowCthjd4KpxVq4M6tdOpXqdVrCU6pnyrBm1w9byNIYsx5AROoinBNW9Ski38IKNjdhtXftBuZgLUPTXe1LZYYYftgaY77EmsH+QRE5BngHq4p1XqTjReR44D7gbGPMypBdDqwVBi6KcFpltHQolUga8JTqAWNMhYi8DlwnIn9qVwKL1THAR8aY4Bg6EYlUGpooIikh1ZdHAk1YS9PE0wr7OeJwChEpweow8xNjzGvtdi/DqordZaxFPpXqN7RKU6mem431f2mpiFwsIhNE5EARuRirfcwf5fzVwGQROVVESkTkDqwOH+25gLkicpCInAj8Cni0s84ysRCRP4vIHSJytIiMtBeIfQKow2rLa3+8F2s9uIXAcyKSH3jYhzwNlAMvi8hxIjJKRKaLyG/tQKlUn9ESnlI9ZIxZJyKTgFuBe7B6LjZjreb8EBBt9pO/YLX3zcdqQ3sBq1v/Ve2OWwx8hdUL0mcf95MeJv8N+z4/wBr2UInV6/JEY8zqCMfnYVXfjgMuaLdPjDF1IjIdKxg/B2RgtSMuAqp6mFalekTH4Sk1AATG4RljTu/rtCg1UGmVplJKqUFBA55SSqlBQas0lVJKDQpawlNKKTUoaMBTSik1KGjAU0opNShowFNKKTUoaMBTSik1KGjAU0opNSj8fxnwCshS1wK3AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,1,figsize=(8, 6))\n",
    "\n",
    "# sns.lineplot(data=df_ds, x='size', y='ratio', hue='type', style='type', marker='x', ax=ax, mec='orange', mfc=\"None\", mew=2)\n",
    "sns.lineplot(data=df_s, x='size', y='ratio', hue='type', style='type', \n",
    "             hue_order=['BA','ER','GRP','WS'],\n",
    "             style_order=['BA','ER','GRP','WS'],\n",
    "             markers=True, markersize=10, dashes=False,\n",
    "             ci=None, \n",
    "             ax=ax)\n",
    "# sns.lineplot(data=df_ds_mis, x='size', y='ratio', hue='type', style='type', markers=True, ax=ax)\n",
    "# sns.lineplot(data=df_ds, x='k', y='ratio', marker='+', ax=axs[1], mec='orange', mfc=\"None\", mew=2)\n",
    "ax.grid(True)\n",
    "ax.set_xticks([100, 150, 200, 250, 300])\n",
    "ax.set_yticks(np.arange(0.8, 1.01,0.02))\n",
    "# ax.set_ylim([0.795, 1.005])\n",
    "ax.set_ylim([0.895, 1.005])\n",
    "# ax.legend(['Barabási–Albert'], fontsize=14)\n",
    "# ax.legend(['Barabási–Albert','Erdős–Rényi','Gaussian Random Partition','Connected Watts-Strogatz small-world'], \n",
    "#           fontsize=14, framealpha=0.3, loc='lower left', title='Graph')\n",
    "ax.legend(fontsize=14, framealpha=0.3, loc='center right', title='Graph')\n",
    "ax.set_xlabel(\"Graph Size\", fontsize=14)\n",
    "ax.set_ylabel(\"Approx. Ratio w.r.t. SPH-NWST\", fontsize=14)\n",
    "fig.set_size_inches(6, 4)\n",
    "fname='./output/nwst_ap_to_greedy.pdf'\n",
    "# pp = PdfPages(fname)\n",
    "plt.tight_layout(pad=0.1)\n",
    "fig.savefig(fname,\n",
    "            dpi=300, #facecolor='w', edgecolor='w',\n",
    "            orientation='portrait',\n",
    "            format='pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>steiner_0</th>\n",
       "      <th>steiner_i</th>\n",
       "      <th>m</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>type</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BA</th>\n",
       "      <td>200.375000</td>\n",
       "      <td>19.625500</td>\n",
       "      <td>0.251237</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>54.503894</td>\n",
       "      <td>59.898244</td>\n",
       "      <td>0.931752</td>\n",
       "      <td>26.720000</td>\n",
       "      <td>24.360000</td>\n",
       "      <td>4.940500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ER</th>\n",
       "      <td>200.625626</td>\n",
       "      <td>19.621121</td>\n",
       "      <td>0.251770</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>13.381993</td>\n",
       "      <td>13.636023</td>\n",
       "      <td>0.985733</td>\n",
       "      <td>14.274775</td>\n",
       "      <td>13.942943</td>\n",
       "      <td>4.943944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GRP</th>\n",
       "      <td>201.400700</td>\n",
       "      <td>19.611806</td>\n",
       "      <td>0.248960</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.241029</td>\n",
       "      <td>8.355667</td>\n",
       "      <td>0.988147</td>\n",
       "      <td>11.247124</td>\n",
       "      <td>11.037519</td>\n",
       "      <td>4.869935</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WS</th>\n",
       "      <td>201.050000</td>\n",
       "      <td>19.684000</td>\n",
       "      <td>0.251287</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>12.292748</td>\n",
       "      <td>12.433117</td>\n",
       "      <td>0.991442</td>\n",
       "      <td>13.430500</td>\n",
       "      <td>13.214500</td>\n",
       "      <td>4.928000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            size          k         p  mwds  mwcds        gcn     greedy  \\\n",
       "type                                                                       \n",
       "BA    200.375000  19.625500  0.251237   NaN    NaN  54.503894  59.898244   \n",
       "ER    200.625626  19.621121  0.251770   NaN    NaN  13.381993  13.636023   \n",
       "GRP   201.400700  19.611806  0.248960   NaN    NaN   8.241029   8.355667   \n",
       "WS    201.050000  19.684000  0.251287   NaN    NaN  12.292748  12.433117   \n",
       "\n",
       "         ratio  steiner_0  steiner_i         m  \n",
       "type                                            \n",
       "BA    0.931752  26.720000  24.360000  4.940500  \n",
       "ER    0.985733  14.274775  13.942943  4.943944  \n",
       "GRP   0.988147  11.247124  11.037519  4.869935  \n",
       "WS    0.991442  13.430500  13.214500  4.928000  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# df_s['ratio'].describe()\n",
    "df_s.groupby(['type']).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Minimum Weight Dominating Set (MWDS) results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>ar_db</th>\n",
       "      <th>ar_gb</th>\n",
       "      <th>mwds_0</th>\n",
       "      <th>mwds_i</th>\n",
       "      <th>vvv</th>\n",
       "      <th>m</th>\n",
       "      <th>Graph</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BA</td>\n",
       "      <td>450</td>\n",
       "      <td>17</td>\n",
       "      <td>0.225536</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19.073828</td>\n",
       "      <td>24.208473</td>\n",
       "      <td>0.787899</td>\n",
       "      <td>0.368194</td>\n",
       "      <td>0.290099</td>\n",
       "      <td>104.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>65.749313</td>\n",
       "      <td>4.0</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BA</td>\n",
       "      <td>400</td>\n",
       "      <td>21</td>\n",
       "      <td>0.191822</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15.095521</td>\n",
       "      <td>17.346281</td>\n",
       "      <td>0.870245</td>\n",
       "      <td>0.435257</td>\n",
       "      <td>0.378781</td>\n",
       "      <td>89.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>39.852931</td>\n",
       "      <td>4.0</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BA</td>\n",
       "      <td>100</td>\n",
       "      <td>18</td>\n",
       "      <td>0.339625</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.323201</td>\n",
       "      <td>1.478952</td>\n",
       "      <td>0.894688</td>\n",
       "      <td>0.691635</td>\n",
       "      <td>0.618798</td>\n",
       "      <td>14.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2.138340</td>\n",
       "      <td>6.0</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BA</td>\n",
       "      <td>350</td>\n",
       "      <td>18</td>\n",
       "      <td>0.284438</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.393639</td>\n",
       "      <td>13.049409</td>\n",
       "      <td>0.873115</td>\n",
       "      <td>0.659495</td>\n",
       "      <td>0.575815</td>\n",
       "      <td>64.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>19.786972</td>\n",
       "      <td>5.0</td>\n",
       "      <td>BA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BA</td>\n",
       "      <td>200</td>\n",
       "      <td>22</td>\n",
       "      <td>0.262928</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.225239</td>\n",
       "      <td>8.030720</td>\n",
       "      <td>0.775178</td>\n",
       "      <td>0.739760</td>\n",
       "      <td>0.573446</td>\n",
       "      <td>44.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>10.855845</td>\n",
       "      <td>6.0</td>\n",
       "      <td>BA</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7995</th>\n",
       "      <td>WS</td>\n",
       "      <td>400</td>\n",
       "      <td>25</td>\n",
       "      <td>0.332019</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.231799</td>\n",
       "      <td>8.231799</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.695276</td>\n",
       "      <td>0.695276</td>\n",
       "      <td>42.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>11.839622</td>\n",
       "      <td>8.0</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7996</th>\n",
       "      <td>WS</td>\n",
       "      <td>100</td>\n",
       "      <td>21</td>\n",
       "      <td>0.227316</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.847273</td>\n",
       "      <td>2.847273</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.026977</td>\n",
       "      <td>2.026977</td>\n",
       "      <td>11.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>1.404689</td>\n",
       "      <td>5.0</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7997</th>\n",
       "      <td>WS</td>\n",
       "      <td>200</td>\n",
       "      <td>13</td>\n",
       "      <td>0.188245</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.442328</td>\n",
       "      <td>7.803584</td>\n",
       "      <td>0.825560</td>\n",
       "      <td>0.840439</td>\n",
       "      <td>0.693833</td>\n",
       "      <td>30.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>9.285125</td>\n",
       "      <td>2.0</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7998</th>\n",
       "      <td>WS</td>\n",
       "      <td>400</td>\n",
       "      <td>27</td>\n",
       "      <td>0.314959</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.219092</td>\n",
       "      <td>9.134770</td>\n",
       "      <td>1.009231</td>\n",
       "      <td>0.885203</td>\n",
       "      <td>0.893374</td>\n",
       "      <td>39.0</td>\n",
       "      <td>39.0</td>\n",
       "      <td>10.319408</td>\n",
       "      <td>9.0</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7999</th>\n",
       "      <td>WS</td>\n",
       "      <td>300</td>\n",
       "      <td>17</td>\n",
       "      <td>0.308333</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.025058</td>\n",
       "      <td>7.937922</td>\n",
       "      <td>1.010977</td>\n",
       "      <td>0.764060</td>\n",
       "      <td>0.772448</td>\n",
       "      <td>38.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>10.389131</td>\n",
       "      <td>5.0</td>\n",
       "      <td>WS</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8000 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     type  size   k         p  mwds  mwcds        gcn     greedy     ratio  \\\n",
       "0      BA   450  17  0.225536   NaN    NaN  19.073828  24.208473  0.787899   \n",
       "1      BA   400  21  0.191822   NaN    NaN  15.095521  17.346281  0.870245   \n",
       "2      BA   100  18  0.339625   NaN    NaN   1.323201   1.478952  0.894688   \n",
       "3      BA   350  18  0.284438   NaN    NaN  11.393639  13.049409  0.873115   \n",
       "4      BA   200  22  0.262928   NaN    NaN   6.225239   8.030720  0.775178   \n",
       "...   ...   ...  ..       ...   ...    ...        ...        ...       ...   \n",
       "7995   WS   400  25  0.332019   NaN    NaN   8.231799   8.231799  1.000000   \n",
       "7996   WS   100  21  0.227316   NaN    NaN   2.847273   2.847273  1.000000   \n",
       "7997   WS   200  13  0.188245   NaN    NaN   6.442328   7.803584  0.825560   \n",
       "7998   WS   400  27  0.314959   NaN    NaN   9.219092   9.134770  1.009231   \n",
       "7999   WS   300  17  0.308333   NaN    NaN   8.025058   7.937922  1.010977   \n",
       "\n",
       "         ar_db     ar_gb  mwds_0  mwds_i        vvv    m Graph  \n",
       "0     0.368194  0.290099   104.0    74.0  65.749313  4.0    BA  \n",
       "1     0.435257  0.378781    89.0    65.0  39.852931  4.0    BA  \n",
       "2     0.691635  0.618798    14.0    12.0   2.138340  6.0    BA  \n",
       "3     0.659495  0.575815    64.0    48.0  19.786972  5.0    BA  \n",
       "4     0.739760  0.573446    44.0    31.0  10.855845  6.0    BA  \n",
       "...        ...       ...     ...     ...        ...  ...   ...  \n",
       "7995  0.695276  0.695276    42.0    42.0  11.839622  8.0    WS  \n",
       "7996  2.026977  2.026977    11.0    11.0   1.404689  5.0    WS  \n",
       "7997  0.840439  0.693833    30.0    27.0   9.285125  2.0    WS  \n",
       "7998  0.885203  0.893374    39.0    39.0  10.319408  9.0    WS  \n",
       "7999  0.764060  0.772448    38.0    38.0  10.389131  5.0    WS  \n",
       "\n",
       "[8000 rows x 16 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mwds_csv_dict = {\n",
    "    \"BA\":'./output/result_GCNDSGDYER_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_BA_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"GRP\":'./output/result_GCNDSGDYER_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_GRP_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"ER\":'./output/result_GCNDSGDYER_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_ER_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "    \"WS\":'./output/result_GCNDSGDYER_deep_ld1_c32_l5_cheb1_diver1_mpy_dpg_policy_WS_Graph_Uniform_GEN21_test1_test_foo.csv',\n",
    "                   }\n",
    "\n",
    "df_ds = pd.DataFrame([])\n",
    "for key in mwds_csv_dict:\n",
    "    mwds_csv = mwds_csv_dict[key]\n",
    "    try:\n",
    "        df_tmp = pd.read_csv(mwds_csv, index_col=0)\n",
    "        df_tmp['m'] = df_tmp['p']*df_tmp['k']\n",
    "        df_tmp['m'] = df_tmp['m'].apply(np.round)\n",
    "        df_tmp['type']=key\n",
    "        df_ds = pd.concat([df_ds, df_tmp], axis=0, ignore_index=True)\n",
    "    except:\n",
    "        pass\n",
    "df_ds['Graph'] = df_ds['type']\n",
    "df_ds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,1,figsize=(8, 6))\n",
    "\n",
    "# sns.lineplot(data=df_ds, x='size', y='ratio', hue='type', style='type', marker='x', ax=ax, mec='orange', mfc=\"None\", mew=2)\n",
    "sns.lineplot(data=df_ds, x='size', y='ratio', hue='Graph', style='Graph', \n",
    "             hue_order=['BA','ER','GRP','WS'],\n",
    "             markers=True, ci=None, markersize=10, dashes=False,\n",
    "             ax=ax)\n",
    "# sns.lineplot(data=df_ds_mis, x='size', y='ratio', hue='type', style='type', markers=True, ax=ax)\n",
    "# sns.lineplot(data=df_ds, x='k', y='ratio', marker='+', ax=axs[1], mec='orange', mfc=\"None\", mew=2)\n",
    "ax.grid(True)\n",
    "ax.set_ylim([0.8, 1.01])\n",
    "# ax.legend(['Erdős–Rényi','Barabási–Albert','Gaussian Random Partition','Connected Watts-Strogatz small-world'], \n",
    "#           fontsize=14, framealpha=0.3)\n",
    "ax.legend(fontsize=14, framealpha=0.3)\n",
    "ax.set_xlabel(\"Graph Size\", fontsize=14)\n",
    "ax.set_ylabel(\"Approx. Ratio w.r.t. Greedy-MWDS\", fontsize=14)\n",
    "fig.set_size_inches(6, 4)\n",
    "fname='./output/mwds_ap_to_greedy.pdf'\n",
    "# pp = PdfPages(fname)\n",
    "plt.tight_layout(pad=0.1)\n",
    "fig.savefig(fname,\n",
    "            dpi=300, #facecolor='w', edgecolor='w',\n",
    "            orientation='portrait',\n",
    "            format='pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>size</th>\n",
       "      <th>k</th>\n",
       "      <th>p</th>\n",
       "      <th>mwds</th>\n",
       "      <th>mwcds</th>\n",
       "      <th>gcn</th>\n",
       "      <th>greedy</th>\n",
       "      <th>ratio</th>\n",
       "      <th>ar_db</th>\n",
       "      <th>ar_gb</th>\n",
       "      <th>mwds_0</th>\n",
       "      <th>mwds_i</th>\n",
       "      <th>vvv</th>\n",
       "      <th>m</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>type</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BA</th>\n",
       "      <td>295.200</td>\n",
       "      <td>19.5645</td>\n",
       "      <td>0.247304</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.768128</td>\n",
       "      <td>13.264352</td>\n",
       "      <td>0.885844</td>\n",
       "      <td>0.467482</td>\n",
       "      <td>0.413254</td>\n",
       "      <td>62.7775</td>\n",
       "      <td>48.3635</td>\n",
       "      <td>30.670492</td>\n",
       "      <td>4.8360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ER</th>\n",
       "      <td>299.675</td>\n",
       "      <td>19.3450</td>\n",
       "      <td>0.250147</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.034774</td>\n",
       "      <td>5.419675</td>\n",
       "      <td>0.932210</td>\n",
       "      <td>0.392283</td>\n",
       "      <td>0.365746</td>\n",
       "      <td>46.2825</td>\n",
       "      <td>42.0390</td>\n",
       "      <td>18.001253</td>\n",
       "      <td>4.8330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GRP</th>\n",
       "      <td>296.200</td>\n",
       "      <td>19.5460</td>\n",
       "      <td>0.250015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.369268</td>\n",
       "      <td>6.626925</td>\n",
       "      <td>0.970221</td>\n",
       "      <td>1.509280</td>\n",
       "      <td>1.449691</td>\n",
       "      <td>27.7030</td>\n",
       "      <td>27.0235</td>\n",
       "      <td>5.297211</td>\n",
       "      <td>4.8995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WS</th>\n",
       "      <td>297.850</td>\n",
       "      <td>19.5660</td>\n",
       "      <td>0.249930</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.429164</td>\n",
       "      <td>8.704170</td>\n",
       "      <td>0.973998</td>\n",
       "      <td>1.006594</td>\n",
       "      <td>0.974317</td>\n",
       "      <td>36.5475</td>\n",
       "      <td>35.8795</td>\n",
       "      <td>11.501232</td>\n",
       "      <td>4.8970</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         size        k         p  mwds  mwcds        gcn     greedy     ratio  \\\n",
       "type                                                                            \n",
       "BA    295.200  19.5645  0.247304   NaN    NaN  11.768128  13.264352  0.885844   \n",
       "ER    299.675  19.3450  0.250147   NaN    NaN   5.034774   5.419675  0.932210   \n",
       "GRP   296.200  19.5460  0.250015   NaN    NaN   6.369268   6.626925  0.970221   \n",
       "WS    297.850  19.5660  0.249930   NaN    NaN   8.429164   8.704170  0.973998   \n",
       "\n",
       "         ar_db     ar_gb   mwds_0   mwds_i        vvv       m  \n",
       "type                                                           \n",
       "BA    0.467482  0.413254  62.7775  48.3635  30.670492  4.8360  \n",
       "ER    0.392283  0.365746  46.2825  42.0390  18.001253  4.8330  \n",
       "GRP   1.509280  1.449691  27.7030  27.0235   5.297211  4.8995  \n",
       "WS    1.006594  0.974317  36.5475  35.8795  11.501232  4.8970  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_ds.groupby(['type']).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Maximum Weight Independent Set (MWIS) results\n",
    "\n",
    "## 1. Local Communication Complexity (MWIS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>p</th>\n",
       "      <th>step_lgs</th>\n",
       "      <th>step_gcn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ER_n100_p0.02_b0_uni.mat</td>\n",
       "      <td>1.002887</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ER_n100_p0.02_b10_uni.mat</td>\n",
       "      <td>0.995925</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ER_n100_p0.02_b11_uni.mat</td>\n",
       "      <td>0.991418</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ER_n100_p0.02_b12_uni.mat</td>\n",
       "      <td>1.010330</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ER_n100_p0.02_b13_uni.mat</td>\n",
       "      <td>0.981121</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>495</th>\n",
       "      <td>ER_n300_p0.067_b5_uni.mat</td>\n",
       "      <td>1.080286</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>496</th>\n",
       "      <td>ER_n300_p0.067_b6_uni.mat</td>\n",
       "      <td>1.018484</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>497</th>\n",
       "      <td>ER_n300_p0.067_b7_uni.mat</td>\n",
       "      <td>1.051308</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>498</th>\n",
       "      <td>ER_n300_p0.067_b8_uni.mat</td>\n",
       "      <td>1.013965</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>499</th>\n",
       "      <td>ER_n300_p0.067_b9_uni.mat</td>\n",
       "      <td>1.129063</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>500 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                          data         p  step_lgs  step_gcn\n",
       "0     ER_n100_p0.02_b0_uni.mat  1.002887         3         2\n",
       "1    ER_n100_p0.02_b10_uni.mat  0.995925         2         2\n",
       "2    ER_n100_p0.02_b11_uni.mat  0.991418         3         2\n",
       "3    ER_n100_p0.02_b12_uni.mat  1.010330         2         2\n",
       "4    ER_n100_p0.02_b13_uni.mat  0.981121         2         2\n",
       "..                         ...       ...       ...       ...\n",
       "495  ER_n300_p0.067_b5_uni.mat  1.080286         4         5\n",
       "496  ER_n300_p0.067_b6_uni.mat  1.018484         4         4\n",
       "497  ER_n300_p0.067_b7_uni.mat  1.051308         4         4\n",
       "498  ER_n300_p0.067_b8_uni.mat  1.013965         4         4\n",
       "499  ER_n300_p0.067_b9_uni.mat  1.129063         5         6\n",
       "\n",
       "[500 rows x 4 columns]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mwis_csv = \"./output/result_GCNTwinAlt_deep_ld1_c32_l3_cheb1_diver1_mwis_dqn_ER_Graph_Uniform_GEN21_test2_lgs_local_complexity.csv\"\n",
    "df_mwis = pd.read_csv(mwis_csv, index_col=0)\n",
    "df_mwis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def extract_Np(filename):\n",
    "    list_para = filename[0:-4].split('_')\n",
    "    N_p = round(float(list_para[2][1:]) * float(list_para[1][1:]), 0)\n",
    "    return N_p\n",
    "\n",
    "\n",
    "def extract_N(filename):\n",
    "    list_para = filename[0:-4].split('_')\n",
    "    N = int(list_para[1][1:])\n",
    "    return N\n",
    "\n",
    "def convertDf(df):\n",
    "    df1 = df.rename(columns={\"file\": 'data', \"ratio\": 'p'})\n",
    "    return df1.set_index('data')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>p</th>\n",
       "      <th>step_lgs</th>\n",
       "      <th>step_gcn</th>\n",
       "      <th>N_p</th>\n",
       "      <th>N</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ER_n100_p0.02_b0_uni.mat</td>\n",
       "      <td>1.002887</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ER_n100_p0.02_b10_uni.mat</td>\n",
       "      <td>0.995925</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ER_n100_p0.02_b11_uni.mat</td>\n",
       "      <td>0.991418</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ER_n100_p0.02_b12_uni.mat</td>\n",
       "      <td>1.010330</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ER_n100_p0.02_b13_uni.mat</td>\n",
       "      <td>0.981121</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</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>495</th>\n",
       "      <td>ER_n300_p0.067_b5_uni.mat</td>\n",
       "      <td>1.080286</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>496</th>\n",
       "      <td>ER_n300_p0.067_b6_uni.mat</td>\n",
       "      <td>1.018484</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>497</th>\n",
       "      <td>ER_n300_p0.067_b7_uni.mat</td>\n",
       "      <td>1.051308</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>498</th>\n",
       "      <td>ER_n300_p0.067_b8_uni.mat</td>\n",
       "      <td>1.013965</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>499</th>\n",
       "      <td>ER_n300_p0.067_b9_uni.mat</td>\n",
       "      <td>1.129063</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>500 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                          data         p  step_lgs  step_gcn   N_p    N\n",
       "0     ER_n100_p0.02_b0_uni.mat  1.002887         3         2   2.0  100\n",
       "1    ER_n100_p0.02_b10_uni.mat  0.995925         2         2   2.0  100\n",
       "2    ER_n100_p0.02_b11_uni.mat  0.991418         3         2   2.0  100\n",
       "3    ER_n100_p0.02_b12_uni.mat  1.010330         2         2   2.0  100\n",
       "4    ER_n100_p0.02_b13_uni.mat  0.981121         2         2   2.0  100\n",
       "..                         ...       ...       ...       ...   ...  ...\n",
       "495  ER_n300_p0.067_b5_uni.mat  1.080286         4         5  20.0  300\n",
       "496  ER_n300_p0.067_b6_uni.mat  1.018484         4         4  20.0  300\n",
       "497  ER_n300_p0.067_b7_uni.mat  1.051308         4         4  20.0  300\n",
       "498  ER_n300_p0.067_b8_uni.mat  1.013965         4         4  20.0  300\n",
       "499  ER_n300_p0.067_b9_uni.mat  1.129063         5         6  20.0  300\n",
       "\n",
       "[500 rows x 6 columns]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_mwis['N_p'] = df_mwis['data'].apply(extract_Np)\n",
    "df_mwis['N'] = df_mwis['data'].apply(extract_N)\n",
    "df_mwis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>p</th>\n",
       "      <th>step_lgs</th>\n",
       "      <th>step_gcn</th>\n",
       "      <th>N_p</th>\n",
       "      <th>N</th>\n",
       "      <th>Solver</th>\n",
       "      <th>step</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ER_n100_p0.02_b0_uni.mat</td>\n",
       "      <td>1.002887</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "      <td>GCNN(3)+LGS</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ER_n100_p0.02_b10_uni.mat</td>\n",
       "      <td>0.995925</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "      <td>GCNN(3)+LGS</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ER_n100_p0.02_b11_uni.mat</td>\n",
       "      <td>0.991418</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "      <td>GCNN(3)+LGS</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ER_n100_p0.02_b12_uni.mat</td>\n",
       "      <td>1.010330</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "      <td>GCNN(3)+LGS</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ER_n100_p0.02_b13_uni.mat</td>\n",
       "      <td>0.981121</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>100</td>\n",
       "      <td>GCNN(3)+LGS</td>\n",
       "      <td>2</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>995</th>\n",
       "      <td>ER_n300_p0.067_b5_uni.mat</td>\n",
       "      <td>1.080286</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "      <td>Vanilla LGS</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>996</th>\n",
       "      <td>ER_n300_p0.067_b6_uni.mat</td>\n",
       "      <td>1.018484</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "      <td>Vanilla LGS</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>997</th>\n",
       "      <td>ER_n300_p0.067_b7_uni.mat</td>\n",
       "      <td>1.051308</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "      <td>Vanilla LGS</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>998</th>\n",
       "      <td>ER_n300_p0.067_b8_uni.mat</td>\n",
       "      <td>1.013965</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "      <td>Vanilla LGS</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999</th>\n",
       "      <td>ER_n300_p0.067_b9_uni.mat</td>\n",
       "      <td>1.129063</td>\n",
       "      <td>5</td>\n",
       "      <td>6</td>\n",
       "      <td>20.0</td>\n",
       "      <td>300</td>\n",
       "      <td>Vanilla LGS</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1000 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                          data         p  step_lgs  step_gcn   N_p    N  \\\n",
       "0     ER_n100_p0.02_b0_uni.mat  1.002887         3         2   2.0  100   \n",
       "1    ER_n100_p0.02_b10_uni.mat  0.995925         2         2   2.0  100   \n",
       "2    ER_n100_p0.02_b11_uni.mat  0.991418         3         2   2.0  100   \n",
       "3    ER_n100_p0.02_b12_uni.mat  1.010330         2         2   2.0  100   \n",
       "4    ER_n100_p0.02_b13_uni.mat  0.981121         2         2   2.0  100   \n",
       "..                         ...       ...       ...       ...   ...  ...   \n",
       "995  ER_n300_p0.067_b5_uni.mat  1.080286         4         5  20.0  300   \n",
       "996  ER_n300_p0.067_b6_uni.mat  1.018484         4         4  20.0  300   \n",
       "997  ER_n300_p0.067_b7_uni.mat  1.051308         4         4  20.0  300   \n",
       "998  ER_n300_p0.067_b8_uni.mat  1.013965         4         4  20.0  300   \n",
       "999  ER_n300_p0.067_b9_uni.mat  1.129063         5         6  20.0  300   \n",
       "\n",
       "          Solver  step  \n",
       "0    GCNN(3)+LGS     2  \n",
       "1    GCNN(3)+LGS     2  \n",
       "2    GCNN(3)+LGS     2  \n",
       "3    GCNN(3)+LGS     2  \n",
       "4    GCNN(3)+LGS     2  \n",
       "..           ...   ...  \n",
       "995  Vanilla LGS     4  \n",
       "996  Vanilla LGS     4  \n",
       "997  Vanilla LGS     4  \n",
       "998  Vanilla LGS     4  \n",
       "999  Vanilla LGS     5  \n",
       "\n",
       "[1000 rows x 8 columns]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1 = df_mwis.copy()\n",
    "df2 = df_mwis.copy()\n",
    "# df1['algo'] = 'GCNN(5)+Dist.Greedy, ' + df1['type']\n",
    "df1['Solver'] = 'GCNN(3)+LGS'\n",
    "df1['step'] = df1['step_gcn']\n",
    "# df2['algo'] = 'Dist.Greedy, ' + df2['type']\n",
    "df2['Solver'] = 'Vanilla LGS'\n",
    "df2['step'] = df2['step_lgs']\n",
    "df_is = pd.concat([df1,df2], ignore_index=True)\n",
    "df_is"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,1,figsize=(8, 6))\n",
    "\n",
    "# sns.lineplot(data=df_ds, x='size', y='ratio', hue='type', style='type', marker='x', ax=ax, mec='orange', mfc=\"None\", mew=2)\n",
    "sns.lineplot(data=df_is, x='N', y='step', hue='Solver', style='Solver', \n",
    "             hue_order=['GCNN(3)+LGS','Vanilla LGS'],\n",
    "             markers=True, ci=None, markersize=10, dashes=False,\n",
    "             ax=ax)\n",
    "# sns.lineplot(data=df_ds_mis, x='size', y='ratio', hue='type', style='type', markers=True, ax=ax)\n",
    "# sns.lineplot(data=df_ds, x='k', y='ratio', marker='+', ax=axs[1], mec='orange', mfc=\"None\", mew=2)\n",
    "ax.grid(True)\n",
    "ax.set_ylim([0, 5])\n",
    "ax.set_xticks([100, 150, 200, 250, 300])\n",
    "# ax.legend(['Erdős–Rényi','Barabási–Albert','Gaussian Random Partition','Connected Watts-Strogatz small-world'], \n",
    "#           fontsize=14, framealpha=0.3)\n",
    "ax.legend(fontsize=14, framealpha=0.3)\n",
    "ax.set_xlabel(\"Graph Size\", fontsize=14)\n",
    "ax.set_ylabel(\"Local Communication Complexity\", fontsize=14)\n",
    "fig.set_size_inches(6, 4)\n",
    "fname='./output/mwis_local_complexity.pdf'\n",
    "# pp = PdfPages(fname)\n",
    "plt.tight_layout(pad=0.1)\n",
    "fig.savefig(fname,\n",
    "            dpi=300, #facecolor='w', edgecolor='w',\n",
    "            orientation='portrait',\n",
    "            format='pdf')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Approximation Ratios (MWIS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "files = [\n",
    "    {\n",
    "        'legend': \"MP(Paschalidis15)\",\n",
    "        'file': 'mp_clique_greedy_ER_Graph_Uniform_GEN21_test2.csv',\n",
    "        'dataset': 'ER_Graph_Uniform_GEN21_test2',\n",
    "        'train': '',\n",
    "        'test': 'ER'\n",
    "    },\n",
    "    {\n",
    "        'legend': \"MP(Paschalidis15)\",\n",
    "        'file': 'mp_clique_greedy_BA_Graph_Uniform_GEN21_test2.csv',\n",
    "        'dataset': 'BA_Graph_Uniform_GEN21_test2',\n",
    "        'train': '',\n",
    "        'test': 'BA'\n",
    "    },\n",
    "    {\n",
    "        'legend': 'IP(Gurobi)',\n",
    "        'file': 'mlp_gurobi_BA_Graph_Uniform_GEN21_test2.csv',\n",
    "        'dataset': 'BA_Graph_Uniform_GEN21_test2',\n",
    "        'train': '',\n",
    "        'test': 'BA'\n",
    "    },\n",
    "    {\n",
    "        'legend': 'IP(Gurobi)',\n",
    "        'file': 'mlp_gurobi_ER_Graph_Uniform_GEN21_test2.csv',\n",
    "        'dataset': 'ER_Graph_Uniform_GEN21_test2',\n",
    "        'train': '',\n",
    "        'test': 'ER'\n",
    "    },\n",
    "    {\n",
    "        'legend': 'GCNN(3)-LGS-adhoc,ER',\n",
    "        'file': 'result_IS4SAT_deep_ld1_c32_l3_cheb1_diver1_mwis_dqn_ER_Graph_Uniform_GEN21_test2_lgs.csv',\n",
    "        'dataset': 'ER_Graph_Uniform_GEN21_test2',\n",
    "        'train': 'ER',\n",
    "        'test': 'ER'\n",
    "    },\n",
    "    {\n",
    "        'legend': 'GCNN(3)-LGS-adhoc,ER',\n",
    "        'file': 'result_IS4SAT_deep_ld1_c32_l3_cheb1_diver1_mwis_dqn_BA_Graph_Uniform_GEN21_test2_lgs.csv',\n",
    "        'dataset': 'BA_Graph_Uniform_GEN21_test2',\n",
    "        'train': 'ER',\n",
    "        'test': 'BA'\n",
    "    },\n",
    "    {\n",
    "        'legend': 'GCNN(3)-LGS-Twin,ER',\n",
    "        'file': 'result_GCNTwinAlt_deep_ld1_c32_l3_cheb1_diver1_mwis_dqn_ER_Graph_Uniform_GEN21_test2_lgs.csv',\n",
    "        'dataset': 'ER_Graph_Uniform_GEN21_test2',\n",
    "        'train': 'ER',\n",
    "        'test': 'ER'\n",
    "    },\n",
    "    {\n",
    "        'legend': 'GCNN(3)-LGS-ZOO,ER',\n",
    "        'file': 'result_GCNZoo4_deep_ld1_c32_l3_cheb1_diver1_mwis_dqn_ER_Graph_Uniform_GEN21_test2_lgs.csv',\n",
    "        'dataset': 'ER_Graph_Uniform_GEN21_test2',\n",
    "        'train': 'ER',\n",
    "        'test': 'ER'\n",
    "    },\n",
    "]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "results = {}\n",
    "data = pd.DataFrame()\n",
    "for test in ['ER']:\n",
    "    csvfile = 'mlp_gurobi_{}_Graph_Uniform_GEN21_test2.csv'.format(test)\n",
    "    df0 = pd.read_csv(os.path.join('./output/', csvfile), index_col='data')\n",
    "    key = 'LP'\n",
    "    df0['LGS(Joo12)'] = 1 / df0['p']\n",
    "    tmp = pd.DataFrame()\n",
    "    tmp['approx_ratio'] = df0['LGS(Joo12)']\n",
    "    tmp['Solver'] = 'LGS(Joo12)'\n",
    "    tmp['test'] = test\n",
    "    tmp.reset_index(inplace=True)\n",
    "    tmp = tmp.rename(columns={\"index\": 'data'})\n",
    "    data = data.append(tmp, ignore_index=True)\n",
    "    for i in range(len(files)):\n",
    "        fdict = files[i]\n",
    "        if test != fdict['test']:\n",
    "            continue\n",
    "        csvfile = os.path.join('./output', fdict['file'])\n",
    "        if not os.path.isfile(csvfile):\n",
    "            continue\n",
    "        df = pd.read_csv(csvfile, index_col=False)\n",
    "        df = convertDf(df)\n",
    "        tmp = pd.DataFrame()\n",
    "        tmp['approx_ratio'] = df['p']*df0['LGS(Joo12)']\n",
    "        tmp['Solver'] = fdict['legend']\n",
    "        tmp['test'] = test\n",
    "        tmp.reset_index(inplace=True)\n",
    "        tmp = tmp.rename(columns={\"index\": 'data'})\n",
    "        data = data.append(tmp, ignore_index=True)\n",
    "\n",
    "data['N_p'] = data['data'].apply(extract_Np)\n",
    "data['N'] = data['data'].apply(extract_N)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def plot_approx_by_size(data, lgddict={}):\n",
    "    fontsize = 14\n",
    "    plotList = [\n",
    "        'GCNN(3)-LGS-adhoc,ER',\n",
    "        'GCNN(3)-LGS-Twin,ER',\n",
    "        'GCNN(3)-LGS-ZOO,ER',\n",
    "        \"MP(Paschalidis15)\",\n",
    "        'LGS(Joo12)',\n",
    "        ]\n",
    "    datap = pd.DataFrame()\n",
    "    for sol in plotList:\n",
    "        tmp = data.loc[data['Solver'] == sol]\n",
    "        tmp = tmp.replace(to_replace=sol, value=sol.replace(',ER',''))\n",
    "        datap = datap.append(tmp, ignore_index=True)\n",
    "    datap = datap.replace(lgddict)\n",
    "\n",
    "    ax = sns.lineplot(x=\"N\", y=\"approx_ratio\", hue=\"Solver\", style='Solver', ci=None,\n",
    "                      linewidth=2.0, markersize=10,\n",
    "                      data=datap, palette=\"tab10\", markers=True, dashes=True)\n",
    "    for key, val in lgddict.items():\n",
    "        datap['Solver'] = datap.Solver.str.replace(key, val, regex=False)\n",
    "        plotList = [a.replace(key, val) for a in plotList]\n",
    "    lgd = [a.replace(',ER','') for a in plotList]\n",
    "    L = ax.legend(lgd, loc='best', ncol=1, fontsize=12,\n",
    "                  fancybox=True, framealpha=0.3, columnspacing=1)\n",
    "\n",
    "    ax.grid()\n",
    "    ax.set_ylim([0.85, 1.005])\n",
    "    # ax.set_yticks(np.arange(0.85, 1.005, 0.05))\n",
    "    ax.set_xticks([100, 150, 200, 250, 300])\n",
    "    ax.set_xlabel('Graph Size', fontsize=fontsize)\n",
    "    ax.set_ylabel('Approx. Ratio w.r.t. Optimal')\n",
    "    for item in ([ax.title, ax.xaxis.label, ax.yaxis.label] +\n",
    "                 ax.get_xticklabels() + ax.get_yticklabels()):\n",
    "        item.set_fontsize(fontsize)\n",
    "    fig = plt.gcf()\n",
    "    fig.set_size_inches(6, 4)\n",
    "    fname='./output/mwis_ap_to_optimal_twin.pdf'\n",
    "    # pp = PdfPages(fname)\n",
    "    plt.tight_layout(pad=0.1)\n",
    "    fig.savefig(fname,\n",
    "                dpi=300, #facecolor='w', edgecolor='w',\n",
    "                orientation='portrait',\n",
    "                format='pdf')\n",
    "    fig.savefig(fname.replace('.pdf','.png'),\n",
    "                dpi=300, #facecolor='w', edgecolor='w',\n",
    "                orientation='portrait',\n",
    "                format='png')\n",
    "    plt.close()\n",
    "    print(datap.groupby(\"Solver\")[\"approx_ratio\"].describe())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                    count      mean       std       min       25%       50%  \\\n",
      "Solver                                                                        \n",
      "GCNN(3)-LGS-Twin    500.0  0.931470  0.048066  0.617983  0.897709  0.941279   \n",
      "GCNN(3)-LGS-ZOO     500.0  0.927229  0.046911  0.603689  0.902223  0.935571   \n",
      "GCNN(3)-LGS-adhoc   500.0  0.931685  0.048746  0.604419  0.899529  0.938812   \n",
      "LGS [Joo12]         500.0  0.897314  0.059907  0.608567  0.857902  0.899586   \n",
      "MP [Paschalidis15]  500.0  0.906907  0.071173  0.603072  0.857545  0.900465   \n",
      "\n",
      "                         75%  max  \n",
      "Solver                             \n",
      "GCNN(3)-LGS-Twin    0.972390  1.0  \n",
      "GCNN(3)-LGS-ZOO     0.962904  1.0  \n",
      "GCNN(3)-LGS-adhoc   0.971741  1.0  \n",
      "LGS [Joo12]         0.945900  1.0  \n",
      "MP [Paschalidis15]  0.982117  1.0  \n"
     ]
    }
   ],
   "source": [
    "lgddict = {'LGS(Joo12)': 'LGS [Joo12]',\n",
    "           'MP(Paschalidis15)': 'MP [Paschalidis15]',\n",
    "           'GCNN(3)': 'GCNN(3)',\n",
    "           }\n",
    "\n",
    "plot_approx_by_size(data, lgddict)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}