View_optimal_parameters.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Use token from .telepythrc.\n"
     ]
    }
   ],
   "source": [
    "import matplotlib as mpl\n",
    "import pandas as pd\n",
    "from pandas.plotting import parallel_coordinates\n",
    "\n",
    "from utils import read_config\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "colors = [c['color'] for c in mpl.rcParams['axes.prop_cycle']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "models = ['baseline', 'scaledsvd', 'hybridsvd', 'scaledhybridsvd', 'lce_rank', 'fm_rank', 'sgd_rank']\n",
    "cs_models = ['baseline_coldstart', 'scaledsvd_coldstart', 'hybridsvd_coldstart', 'scaledhybridsvd_coldstart', 'lce_coldstart_rank', 'fm_coldstart_rank']\n",
    "data_labels = ['ML1M', 'ML10M', 'BX', 'AMZvg', 'AMZe', 'YaMus']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Standard case"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ML1M\n",
      "\t {'PureSVD': {'rank': 30}}\n",
      "\t {'PureSVDs': {'col_scaling': 0.4, 'rank': 75}}\n",
      "\t {'HybridSVD': {'features_weight': 0.9, 'rank': 50}}\n",
      "\t {'HybridSVDs': {'col_scaling': 0.4, 'features_weight': 0.7, 'rank': 100}}\n",
      "\t {'LCE': {'alpha': 0.1, 'beta': 0.1, 'rank': 150}}\n",
      "\t {'FM': {'regularization': 3e-08, 'linear_regularization': 1e-10, 'rank': 750}}\n",
      "\t {'SGD': {'regularization': 3e-06, 'linear_regularization': 1e-10, 'rank': 1000}}\n",
      "ML10M\n",
      "\t {'PureSVD': {'rank': 30}}\n",
      "\t {'PureSVDs': {'col_scaling': 0.2, 'rank': 1000}}\n",
      "\t {'HybridSVD': {'features_weight': 0.1, 'rank': 30}}\n",
      "\t {'HybridSVDs': {'col_scaling': 0.2, 'features_weight': 0.1, 'rank': 1000}}\n",
      "\t {'LCE': {'alpha': 0.3, 'beta': 0.1, 'rank': 1000}}\n",
      "\t {'FM': {'regularization': 1e-06, 'linear_regularization': 3e-07, 'rank': 750}}\n",
      "\t {'SGD': {'regularization': 1e-06, 'linear_regularization': 1e-07, 'rank': 750}}\n",
      "BX\n",
      "\t {'PureSVD': {'rank': 1250}}\n",
      "\t {'PureSVDs': {'col_scaling': 0.4, 'rank': 1000}}\n",
      "\t {'HybridSVD': {'features_weight': 0.3, 'rank': 1250}}\n",
      "\t {'HybridSVDs': {'col_scaling': 0.4, 'features_weight': 0.5, 'rank': 1750}}\n",
      "\t {'LCE': {'alpha': 0.1, 'beta': 0.0, 'rank': 1500}}\n",
      "\t {'FM': {'regularization': 1e-06, 'linear_regularization': 3e-08, 'rank': 750}}\n",
      "\t {'SGD': {'regularization': 3e-06, 'linear_regularization': 1e-08, 'rank': 1500}}\n",
      "AMZvg\n",
      "\t {'PureSVD': {'rank': 150}}\n",
      "\t {'PureSVDs': {'col_scaling': 0.4, 'rank': 300}}\n",
      "\t {'HybridSVD': {'features_weight': 0.1, 'rank': 150}}\n",
      "\t {'HybridSVDs': {'col_scaling': 0.4, 'features_weight': 0.3, 'rank': 300}}\n",
      "\t {'LCE': {'alpha': 0.1, 'beta': 0.3, 'rank': 2500}}\n",
      "\t {'FM': {'regularization': 3e-06, 'linear_regularization': 3e-06, 'rank': 2000}}\n",
      "\t {'SGD': {'regularization': 3e-06, 'linear_regularization': 1e-07, 'rank': 1500}}\n",
      "AMZe\n",
      "\t {'PureSVD': {'rank': 1000}}\n",
      "\t {'PureSVDs': {'col_scaling': 0.2, 'rank': 3000}}\n",
      "\t {'HybridSVD': {'features_weight': 0.1, 'rank': 1500}}\n",
      "\t {'HybridSVDs': {'col_scaling': 0.2, 'features_weight': 0.1, 'rank': 3000}}\n",
      "\t {'LCE': {'alpha': 0.1, 'beta': 0.0, 'rank': 100}}\n",
      "\t {'FM': {'regularization': 1e-06, 'linear_regularization': 1e-08, 'rank': 1750}}\n",
      "\t {'SGD': {'regularization': 1e-06, 'linear_regularization': 1e-08, 'rank': 1500}}\n",
      "YaMus\n",
      "\t {'PureSVD': {'rank': 3000}}\n",
      "\t {'PureSVDs': {'col_scaling': 0.4, 'rank': 3000}}\n",
      "\t {'HybridSVD': {'features_weight': 0.7, 'rank': 3000}}\n",
      "\t {'HybridSVDs': {'col_scaling': 0.4, 'features_weight': 0.1, 'rank': 3000}}\n",
      "\t {'LCE': {'alpha': 0.7, 'beta': 0.3, 'rank': 3000}}\n",
      "\t {'FM': {'regularization': 1e-09, 'linear_regularization': 1e-10, 'rank': 1}}\n",
      "\t {'SGD': {'regularization': 3e-08, 'linear_regularization': 3e-07, 'rank': 3000}}\n"
     ]
    }
   ],
   "source": [
    "for label in data_labels:\n",
    "    print(label)\n",
    "    for model in models:\n",
    "        try:\n",
    "            config = read_config(model, label)\n",
    "        except FileNotFoundError:\n",
    "            print(f'\\t !No data for {model}')\n",
    "        else:\n",
    "            if 'scaledhybridsvd' in model:\n",
    "                params =  read_config('scaledsvd', label)\n",
    "                config['HybridSVDs'] = {**{k:v for k,v in params['PureSVDs'].items() if k=='col_scaling'},\n",
    "                                        **config['HybridSVDs']}            \n",
    "            if 'lce' in model:\n",
    "                params = read_config('lce_param', label)\n",
    "                config['LCE'] = {**{k:v for k,v in params['LCE'].items() if k in ['alpha', 'beta']}, **config['LCE']}\n",
    "            if 'fm' in model:\n",
    "                params = read_config('fm_param', label)\n",
    "                config['FM'] = {**{k:v for k,v in params['FM'].items() if k.endswith('regularization')}, **config['FM']}\n",
    "            if 'sgd' in model:\n",
    "                params = read_config('sgd_param', label)\n",
    "                config['SGD'] = {**{k:v for k,v in params['SGD'].items() if k.endswith('regularization')}, **config['SGD']}\n",
    "            print('\\t', config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## model rank comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "ranks = []\n",
    "for label in data_labels:\n",
    "    for model in models:\n",
    "        try:\n",
    "            config = read_config(model, label)\n",
    "        except FileNotFoundError:\n",
    "            pass\n",
    "        else:\n",
    "            for name, params in config.items():\n",
    "                ranks.append((name, label, params['rank']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>model</th>\n",
       "      <th>data</th>\n",
       "      <th>rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PureSVDs</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HybridSVD</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HybridSVDs</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LCE</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        model  data  rank\n",
       "0     PureSVD  ML1M    30\n",
       "1    PureSVDs  ML1M    75\n",
       "2   HybridSVD  ML1M    50\n",
       "3  HybridSVDs  ML1M   100\n",
       "4         LCE  ML1M   150"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_config = pd.DataFrame(ranks, columns=['model', 'data', 'rank'])\n",
    "opt_config.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>data</th>\n",
       "      <th>AMZe</th>\n",
       "      <th>AMZvg</th>\n",
       "      <th>BX</th>\n",
       "      <th>ML10M</th>\n",
       "      <th>ML1M</th>\n",
       "      <th>YaMus</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</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>FM</th>\n",
       "      <td>1750</td>\n",
       "      <td>2000</td>\n",
       "      <td>750</td>\n",
       "      <td>750</td>\n",
       "      <td>750</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HybridSVD</th>\n",
       "      <td>1500</td>\n",
       "      <td>150</td>\n",
       "      <td>1250</td>\n",
       "      <td>30</td>\n",
       "      <td>50</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HybridSVDs</th>\n",
       "      <td>3000</td>\n",
       "      <td>300</td>\n",
       "      <td>1750</td>\n",
       "      <td>1000</td>\n",
       "      <td>100</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LCE</th>\n",
       "      <td>100</td>\n",
       "      <td>2500</td>\n",
       "      <td>1500</td>\n",
       "      <td>1000</td>\n",
       "      <td>150</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PureSVD</th>\n",
       "      <td>1000</td>\n",
       "      <td>150</td>\n",
       "      <td>1250</td>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            rank                             \n",
       "data        AMZe AMZvg    BX ML10M ML1M YaMus\n",
       "model                                        \n",
       "FM          1750  2000   750   750  750     1\n",
       "HybridSVD   1500   150  1250    30   50  3000\n",
       "HybridSVDs  3000   300  1750  1000  100  3000\n",
       "LCE          100  2500  1500  1000  150  3000\n",
       "PureSVD     1000   150  1250    30   30  3000"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_config.set_index(['model','data']).sort_index().unstack('data').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c18afc18>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(opt_config.set_index(['model','data']).sort_index().unstack('data').loc[:, 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'model', color=colors)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## svd hybridization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>model</th>\n",
       "      <th>data</th>\n",
       "      <th>rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PureSVDs</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HybridSVD</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HybridSVDs</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>ML10M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        model   data  rank\n",
       "0     PureSVD   ML1M    30\n",
       "1    PureSVDs   ML1M    75\n",
       "2   HybridSVD   ML1M    50\n",
       "3  HybridSVDs   ML1M   100\n",
       "7     PureSVD  ML10M    30"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_svd_config = opt_config.loc[opt_config['model'].str.contains('SVD')]\n",
    "opt_svd_config.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>model</th>\n",
       "      <th colspan=\"6\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>data</th>\n",
       "      <th></th>\n",
       "      <th>AMZe</th>\n",
       "      <th>AMZvg</th>\n",
       "      <th>BX</th>\n",
       "      <th>ML10M</th>\n",
       "      <th>ML1M</th>\n",
       "      <th>YaMus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HybridSVD</td>\n",
       "      <td>1500</td>\n",
       "      <td>150</td>\n",
       "      <td>1250</td>\n",
       "      <td>30</td>\n",
       "      <td>50</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HybridSVDs</td>\n",
       "      <td>3000</td>\n",
       "      <td>300</td>\n",
       "      <td>1750</td>\n",
       "      <td>1000</td>\n",
       "      <td>100</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>1000</td>\n",
       "      <td>150</td>\n",
       "      <td>1250</td>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PureSVDs</td>\n",
       "      <td>3000</td>\n",
       "      <td>300</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>75</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           model  rank                             \n",
       "data              AMZe AMZvg    BX ML10M ML1M YaMus\n",
       "0      HybridSVD  1500   150  1250    30   50  3000\n",
       "1     HybridSVDs  3000   300  1750  1000  100  3000\n",
       "2        PureSVD  1000   150  1250    30   30  3000\n",
       "3       PureSVDs  3000   300  1000  1000   75  3000"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_svd_config.set_index(['model','data']).sort_index().unstack('data').reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c1c47da0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(opt_svd_config.set_index(['model','data']).sort_index().unstack('data').loc[:, 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'model', color=colors)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## svd scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaling = []\n",
    "for label in data_labels:\n",
    "    for model in models:\n",
    "        try:\n",
    "            config = read_config(model, label)\n",
    "        except FileNotFoundError:\n",
    "            pass\n",
    "        else:\n",
    "            for name, params in config.items():\n",
    "                if name.endswith('SVD'):\n",
    "                    scaling.append((name, False, label, params['rank']))\n",
    "                if name.endswith('SVDs'):\n",
    "                    scaling.append((name[:-1], True, label, params['rank']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>model</th>\n",
       "      <th>scaling</th>\n",
       "      <th>data</th>\n",
       "      <th>rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>False</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>True</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HybridSVD</td>\n",
       "      <td>False</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HybridSVD</td>\n",
       "      <td>True</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PureSVD</td>\n",
       "      <td>False</td>\n",
       "      <td>ML10M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       model  scaling   data  rank\n",
       "0    PureSVD    False   ML1M    30\n",
       "1    PureSVD     True   ML1M    75\n",
       "2  HybridSVD    False   ML1M    50\n",
       "3  HybridSVD     True   ML1M   100\n",
       "4    PureSVD    False  ML10M    30"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.DataFrame(scaling, columns=['model','scaling', 'data', 'rank'])\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>AMZe</th>\n",
       "      <th>AMZvg</th>\n",
       "      <th>BX</th>\n",
       "      <th>ML10M</th>\n",
       "      <th>ML1M</th>\n",
       "      <th>YaMus</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</th>\n",
       "      <th>scaling</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">HybridSVD</th>\n",
       "      <th>False</th>\n",
       "      <td>1500</td>\n",
       "      <td>150</td>\n",
       "      <td>1250</td>\n",
       "      <td>30</td>\n",
       "      <td>50</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>3000</td>\n",
       "      <td>300</td>\n",
       "      <td>1750</td>\n",
       "      <td>1000</td>\n",
       "      <td>100</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">PureSVD</th>\n",
       "      <th>False</th>\n",
       "      <td>1000</td>\n",
       "      <td>150</td>\n",
       "      <td>1250</td>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>3000</td>\n",
       "      <td>300</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>75</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   rank                             \n",
       "data               AMZe AMZvg    BX ML10M ML1M YaMus\n",
       "model     scaling                                   \n",
       "HybridSVD False    1500   150  1250    30   50  3000\n",
       "          True     3000   300  1750  1000  100  3000\n",
       "PureSVD   False    1000   150  1250    30   30  3000\n",
       "          True     3000   300  1000  1000   75  3000"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.set_index(['model','scaling','data']).sort_index().unstack('data').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c2cc6240>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('data').loc['HybridSVD', 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'scaling', color=colors)\n",
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('data').loc['PureSVD', 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'scaling', color=colors, alpha=0.5)"
   ]
  },
  {
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>scaling</th>\n",
       "      <th>False</th>\n",
       "      <th>True</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</th>\n",
       "      <th>data</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">HybridSVD</th>\n",
       "      <th>AMZe</th>\n",
       "      <td>1500</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMZvg</th>\n",
       "      <td>150</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BX</th>\n",
       "      <td>1250</td>\n",
       "      <td>1750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML10M</th>\n",
       "      <td>30</td>\n",
       "      <td>1000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML1M</th>\n",
       "      <td>50</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 rank      \n",
       "scaling         False True \n",
       "model     data             \n",
       "HybridSVD AMZe   1500  3000\n",
       "          AMZvg   150   300\n",
       "          BX     1250  1750\n",
       "          ML10M    30  1000\n",
       "          ML1M     50   100"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.set_index(['model','scaling','data']).sort_index().unstack('scaling').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c2d29be0>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('scaling').loc['HybridSVD', 'rank'].reset_index().dropna(),\n",
    "                     'data', color=colors)\n",
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('scaling').loc['PureSVD', 'rank'].reset_index().dropna(),\n",
    "                     'data', color=colors, alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>model</th>\n",
       "      <th>HybridSVD</th>\n",
       "      <th>PureSVD</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scaling</th>\n",
       "      <th>data</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"6\" valign=\"top\">False</th>\n",
       "      <th>AMZe</th>\n",
       "      <td>1500</td>\n",
       "      <td>1000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMZvg</th>\n",
       "      <td>150</td>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BX</th>\n",
       "      <td>1250</td>\n",
       "      <td>1250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML10M</th>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML1M</th>\n",
       "      <td>50</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YaMus</th>\n",
       "      <td>3000</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"6\" valign=\"top\">True</th>\n",
       "      <th>AMZe</th>\n",
       "      <td>3000</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMZvg</th>\n",
       "      <td>300</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BX</th>\n",
       "      <td>1750</td>\n",
       "      <td>1000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML10M</th>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML1M</th>\n",
       "      <td>100</td>\n",
       "      <td>75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YaMus</th>\n",
       "      <td>3000</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   rank        \n",
       "model         HybridSVD PureSVD\n",
       "scaling data                   \n",
       "False   AMZe       1500    1000\n",
       "        AMZvg       150     150\n",
       "        BX         1250    1250\n",
       "        ML10M        30      30\n",
       "        ML1M         50      30\n",
       "        YaMus      3000    3000\n",
       "True    AMZe       3000    3000\n",
       "        AMZvg       300     300\n",
       "        BX         1750    1000\n",
       "        ML10M      1000    1000\n",
       "        ML1M        100      75\n",
       "        YaMus      3000    3000"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.set_index(['model','scaling','data']).sort_index().unstack('model')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cold start"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ML1M(CS)\n",
      "\t {'PureSVD(cs)': {'rank': 100}}\n",
      "\t {'PureSVDs(cs)': {'col_scaling': 0.6, 'rank': 30}}\n",
      "\t {'HybridSVD(cs)': {'features_weight': 0.7, 'rank': 5}}\n",
      "\t {'HybridSVDs(cs)': {'col_scaling': 0.6, 'features_weight': 0.7, 'rank': 5}}\n",
      "\t {'LCE(cs)': {'alpha': 0.9, 'beta': 0.0, 'rank': 1}}\n",
      "\t {'FM(cs)': {'regularization': 1e-10, 'linear_regularization': 1e-06, 'rank': 5}}\n",
      "ML10M(CS)\n",
      "\t {'PureSVD(cs)': {'rank': 400}}\n",
      "\t {'PureSVDs(cs)': {'col_scaling': 0.2, 'rank': 60}}\n",
      "\t {'HybridSVD(cs)': {'features_weight': 0.9, 'rank': 60}}\n",
      "\t {'HybridSVDs(cs)': {'col_scaling': 0.2, 'features_weight': 0.9, 'rank': 50}}\n",
      "\t {'LCE(cs)': {'alpha': 0.3, 'beta': 0.0, 'rank': 15}}\n",
      "\t {'FM(cs)': {'regularization': 3e-06, 'linear_regularization': 1e-06, 'rank': 75}}\n",
      "BX(CS)\n",
      "\t {'PureSVD(cs)': {'rank': 400}}\n",
      "\t {'PureSVDs(cs)': {'col_scaling': 0.2, 'rank': 400}}\n",
      "\t {'HybridSVD(cs)': {'features_weight': 0.9, 'rank': 750}}\n",
      "\t {'HybridSVDs(cs)': {'col_scaling': 0.2, 'features_weight': 0.9, 'rank': 500}}\n",
      "\t {'LCE(cs)': {'alpha': 0.1, 'beta': 0.0, 'rank': 75}}\n",
      "\t {'FM(cs)': {'regularization': 3e-06, 'linear_regularization': 3e-07, 'rank': 150}}\n",
      "AMZvg(CS)\n",
      "\t {'PureSVD(cs)': {'rank': 3000}}\n",
      "\t {'PureSVDs(cs)': {'col_scaling': 0.6, 'rank': 1500}}\n",
      "\t {'HybridSVD(cs)': {'features_weight': 0.1, 'rank': 2000}}\n",
      "\t {'HybridSVDs(cs)': {'col_scaling': 0.6, 'features_weight': 0.3, 'rank': 500}}\n",
      "\t {'LCE(cs)': {'alpha': 0.3, 'beta': 0.05, 'rank': 2000}}\n",
      "\t {'FM(cs)': {'regularization': 1e-09, 'linear_regularization': 1e-10, 'rank': 50}}\n",
      "AMZe(CS)\n",
      "\t {'PureSVD(cs)': {'rank': 2500}}\n",
      "\t {'PureSVDs(cs)': {'col_scaling': 0.4, 'rank': 300}}\n",
      "\t {'HybridSVD(cs)': {'features_weight': 0.5, 'rank': 1750}}\n",
      "\t {'HybridSVDs(cs)': {'col_scaling': 0.4, 'features_weight': 0.9, 'rank': 750}}\n",
      "\t {'LCE(cs)': {'alpha': 0.3, 'beta': 0.3, 'rank': 3000}}\n",
      "\t {'FM(cs)': {'regularization': 3e-06, 'linear_regularization': 3e-08, 'rank': 1000}}\n",
      "YaMus(CS)\n",
      "\t {'PureSVD(cs)': {'rank': 2000}}\n",
      "\t {'PureSVDs(cs)': {'col_scaling': 0.2, 'rank': 1250}}\n",
      "\t {'HybridSVD(cs)': {'features_weight': 0.9, 'rank': 2500}}\n",
      "\t {'HybridSVDs(cs)': {'col_scaling': 0.2, 'features_weight': 0.9, 'rank': 2000}}\n",
      "\t {'LCE(cs)': {'alpha': 0.1, 'beta': 0.1, 'rank': 500}}\n",
      "\t !No data for fm_coldstart_rank\n"
     ]
    }
   ],
   "source": [
    "for label in data_labels:\n",
    "    print(label + '(CS)')\n",
    "    for model in cs_models:\n",
    "        try:\n",
    "            config = read_config(model, label)\n",
    "        except FileNotFoundError:\n",
    "            print(f'\\t !No data for {model}')\n",
    "        else:\n",
    "            if 'lce' in model:\n",
    "                params = read_config('lce_coldstart_param', label)\n",
    "                config['LCE(cs)'] = {**{k:v for k,v in params['LCE(cs)'].items() if k in ['alpha', 'beta']},\n",
    "                                     **config['LCE(cs)']}\n",
    "            if 'fm' in model:\n",
    "                params = read_config('fm_coldstart_param', label)\n",
    "                config['FM(cs)'] = {**{k:v for k,v in params['FM(cs)'].items() if k.endswith('regularization')},\n",
    "                                    **config['FM(cs)']}\n",
    "            if 'scaledhybridsvd' in model:\n",
    "                params =  read_config('scaledsvd_coldstart', label)\n",
    "                config['HybridSVDs(cs)'] = {**{k:v for k,v in params['PureSVDs(cs)'].items() if k=='col_scaling'},\n",
    "                                            **config['HybridSVDs(cs)']}\n",
    "            print('\\t', config)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## model rank comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "ranks = []\n",
    "for label in data_labels:\n",
    "    for model in cs_models:\n",
    "        try:\n",
    "            config = read_config(model, label)\n",
    "        except FileNotFoundError:\n",
    "            pass\n",
    "        else:\n",
    "            for name, params in config.items():\n",
    "                ranks.append((name, label, params['rank']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>model</th>\n",
       "      <th>data</th>\n",
       "      <th>rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PureSVDs(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HybridSVD(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HybridSVDs(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LCE(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            model  data  rank\n",
       "0     PureSVD(cs)  ML1M   100\n",
       "1    PureSVDs(cs)  ML1M    30\n",
       "2   HybridSVD(cs)  ML1M     5\n",
       "3  HybridSVDs(cs)  ML1M     5\n",
       "4         LCE(cs)  ML1M     1"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_config = pd.DataFrame(ranks, columns=['model', 'data', 'rank'])\n",
    "opt_config.head()"
   ]
  },
  {
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>data</th>\n",
       "      <th>AMZe</th>\n",
       "      <th>AMZvg</th>\n",
       "      <th>BX</th>\n",
       "      <th>ML10M</th>\n",
       "      <th>ML1M</th>\n",
       "      <th>YaMus</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</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>FM(cs)</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>150.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HybridSVD(cs)</th>\n",
       "      <td>1750.0</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>750.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2500.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HybridSVDs(cs)</th>\n",
       "      <td>750.0</td>\n",
       "      <td>500.0</td>\n",
       "      <td>500.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LCE(cs)</th>\n",
       "      <td>3000.0</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>500.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PureSVD(cs)</th>\n",
       "      <td>2500.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>2000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  rank                                     \n",
       "data              AMZe   AMZvg     BX  ML10M   ML1M   YaMus\n",
       "model                                                      \n",
       "FM(cs)          1000.0    50.0  150.0   75.0    5.0     NaN\n",
       "HybridSVD(cs)   1750.0  2000.0  750.0   60.0    5.0  2500.0\n",
       "HybridSVDs(cs)   750.0   500.0  500.0   50.0    5.0  2000.0\n",
       "LCE(cs)         3000.0  2000.0   75.0   15.0    1.0   500.0\n",
       "PureSVD(cs)     2500.0  3000.0  400.0  400.0  100.0  2000.0"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_config.set_index(['model','data']).sort_index().unstack('data').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c2e1d7b8>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(opt_config.set_index(['model','data']).sort_index().unstack('data').loc[:, 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'model', color=colors)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## svd hybridization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>model</th>\n",
       "      <th>data</th>\n",
       "      <th>rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PureSVDs(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HybridSVD(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HybridSVDs(cs)</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>ML10M</td>\n",
       "      <td>400</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            model   data  rank\n",
       "0     PureSVD(cs)   ML1M   100\n",
       "1    PureSVDs(cs)   ML1M    30\n",
       "2   HybridSVD(cs)   ML1M     5\n",
       "3  HybridSVDs(cs)   ML1M     5\n",
       "6     PureSVD(cs)  ML10M   400"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_svd_config = opt_config.loc[opt_config['model'].str.contains('SVD')]\n",
    "opt_svd_config.head()"
   ]
  },
  {
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>model</th>\n",
       "      <th colspan=\"6\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>data</th>\n",
       "      <th></th>\n",
       "      <th>AMZe</th>\n",
       "      <th>AMZvg</th>\n",
       "      <th>BX</th>\n",
       "      <th>ML10M</th>\n",
       "      <th>ML1M</th>\n",
       "      <th>YaMus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HybridSVD(cs)</td>\n",
       "      <td>1750</td>\n",
       "      <td>2000</td>\n",
       "      <td>750</td>\n",
       "      <td>60</td>\n",
       "      <td>5</td>\n",
       "      <td>2500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HybridSVDs(cs)</td>\n",
       "      <td>750</td>\n",
       "      <td>500</td>\n",
       "      <td>500</td>\n",
       "      <td>50</td>\n",
       "      <td>5</td>\n",
       "      <td>2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>2500</td>\n",
       "      <td>3000</td>\n",
       "      <td>400</td>\n",
       "      <td>400</td>\n",
       "      <td>100</td>\n",
       "      <td>2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PureSVDs(cs)</td>\n",
       "      <td>300</td>\n",
       "      <td>1500</td>\n",
       "      <td>400</td>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>1250</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               model  rank                            \n",
       "data                  AMZe AMZvg   BX ML10M ML1M YaMus\n",
       "0      HybridSVD(cs)  1750  2000  750    60    5  2500\n",
       "1     HybridSVDs(cs)   750   500  500    50    5  2000\n",
       "2        PureSVD(cs)  2500  3000  400   400  100  2000\n",
       "3       PureSVDs(cs)   300  1500  400    60   30  1250"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opt_svd_config.set_index(['model','data']).sort_index().unstack('data').reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c2e71b70>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(opt_svd_config.set_index(['model','data']).sort_index().unstack('data').loc[:, 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'model', color=colors)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## svd scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaling = []\n",
    "for label in data_labels:\n",
    "    for model in cs_models:\n",
    "        try:\n",
    "            config = read_config(model, label)\n",
    "        except FileNotFoundError:\n",
    "            pass\n",
    "        else:\n",
    "            for name, params in config.items():\n",
    "                if name.endswith('SVD(cs)'):\n",
    "                    scaling.append((name, False, label, params['rank']))\n",
    "                if name.endswith('SVDs(cs)'):\n",
    "                    scaling.append((name[:-5]+'(cs)', True, label, params['rank']))"
   ]
  },
  {
   "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>model</th>\n",
       "      <th>scaling</th>\n",
       "      <th>data</th>\n",
       "      <th>rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>False</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>True</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HybridSVD(cs)</td>\n",
       "      <td>False</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HybridSVD(cs)</td>\n",
       "      <td>True</td>\n",
       "      <td>ML1M</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PureSVD(cs)</td>\n",
       "      <td>False</td>\n",
       "      <td>ML10M</td>\n",
       "      <td>400</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           model  scaling   data  rank\n",
       "0    PureSVD(cs)    False   ML1M   100\n",
       "1    PureSVD(cs)     True   ML1M    30\n",
       "2  HybridSVD(cs)    False   ML1M     5\n",
       "3  HybridSVD(cs)     True   ML1M     5\n",
       "4    PureSVD(cs)    False  ML10M   400"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.DataFrame(scaling, columns=['model','scaling', 'data', 'rank'])\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas.plotting import parallel_coordinates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>data</th>\n",
       "      <th>AMZe</th>\n",
       "      <th>AMZvg</th>\n",
       "      <th>BX</th>\n",
       "      <th>ML10M</th>\n",
       "      <th>ML1M</th>\n",
       "      <th>YaMus</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</th>\n",
       "      <th>scaling</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">HybridSVD(cs)</th>\n",
       "      <th>False</th>\n",
       "      <td>1750</td>\n",
       "      <td>2000</td>\n",
       "      <td>750</td>\n",
       "      <td>60</td>\n",
       "      <td>5</td>\n",
       "      <td>2500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>750</td>\n",
       "      <td>500</td>\n",
       "      <td>500</td>\n",
       "      <td>50</td>\n",
       "      <td>5</td>\n",
       "      <td>2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">PureSVD(cs)</th>\n",
       "      <th>False</th>\n",
       "      <td>2500</td>\n",
       "      <td>3000</td>\n",
       "      <td>400</td>\n",
       "      <td>400</td>\n",
       "      <td>100</td>\n",
       "      <td>2000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>300</td>\n",
       "      <td>1500</td>\n",
       "      <td>400</td>\n",
       "      <td>60</td>\n",
       "      <td>30</td>\n",
       "      <td>1250</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       rank                            \n",
       "data                   AMZe AMZvg   BX ML10M ML1M YaMus\n",
       "model         scaling                                  \n",
       "HybridSVD(cs) False    1750  2000  750    60    5  2500\n",
       "              True      750   500  500    50    5  2000\n",
       "PureSVD(cs)   False    2500  3000  400   400  100  2000\n",
       "              True      300  1500  400    60   30  1250"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.set_index(['model','scaling','data']).sort_index().unstack('data').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c101c828>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAD8CAYAAACLrvgBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzsvXl8VdXV///e92ae5xCSQBIIY5ghIKjV4oDWOk84YK3WqVY7WGv7PE+rfbTWp4O/2lL9auuICorzLFpHkABhCAmBJGQg8zwPd9y/P/ZNCBIgwx3Jfr9e95Xcfc85e52Tm/M5e6+11xJSSjQajUajGQ4GTxug0Wg0Gt9Bi4ZGo9Foho0WDY1Go9EMGy0aGo1Goxk2WjQ0Go1GM2y0aGg0Go1m2GjR0Gg0Gs2w0aKh0Wg0mmGjRUOj0Wg0w8bP0wYcDyGEXLRokafN8Apyc3PR10Khr8Vh9LU4jL4Wh8nNzW2SUsa74tjCm9OICCGkN9vnToQQ6Guh0NfiMPpaHEZfi8MIIXKllItdcWw9PaXRaDSaYaNFQ6PRaDTDRouGRqPRaIaNVzvCNRqNZrhER0dTVlZGX1+fp01xG0FBQaSkpODv7++2PrVoaDSak4L777+f8PBw0tLSEEJ42hyXI6WkubmZqqoq0tPT3dbvCaenhBBBQohtQog9QogCIcQDjvZ0IUSOEKJYCLFBCBHgaA90vC9xfJ426Fi/drQfEEKc66qT0mg044+pU6cSGxs7LgQDVLRYbGys20dWw/FpmIDvSinnAfOBVUKIZcAjwKNSykygFbjJsf1NQKuUcirwqGM7hBCzgKuB2cAq4J9CCKMzT0aj0YxfDAbDuBGMfjxxvicUDanocrz1d7wk8F1go6P9OeBix+8XOd7j+HylUGd2EbBeSmmSUpYBJUC2U85CoxmH9JitAPSabR62RDOeGFb0lBDCKITYDTQAm4CDQJuU0urYpApIdvyeDFQCOD5vB2IHtw+xz+C+bhFC7BBC7Bj56Wg044ec0hYAnt5cxuaSJi0eXoDRaGT+/PkDr/Ly8mNuW15eTlZWlvuMcxLDcoRLKW3AfCFEFPAGMHOozRw/hxovyeO0f7uvJ4EnQa0IH459Gs14w2KzU1jXAUB6XCjby1vYXdnG3JRIFk6KJjRQx7h4guDgYHbv3u1pM1zKiNZpSCnbgM+BZUCUEKL/m5kC1Dh+rwJSARyfRwItg9uH2Eej0YyAkoYuTBY7AOfPSeL6ZZOZEh9KbkUrz2wu44uiRrpM1hMcReMOysvLOe2001i4cCELFy5ky5YtR21TUFBAdnY28+fPZ+7cuRQXFwOwbt26gfZbb70Vm83zo8kTPo4IIeIBi5SyTQgRDJyFcm5/BlwOrAduAN5y7PK24/03js//I6WUQoi3gZeEEH8FJgKZwDYnn49GMy7Ir24nKuRwbH5sWCCrspJYmh5LTlkLuw+1kVfZRlZKJIsnRxMe5L44fm/ggXcK2FfT4dRjzpoYwe++P/u42/T29jJ//nwA0tPTeeONN0hISGDTpk0EBQVRXFzM6tWr2bHjyNn3J554grvvvptrr70Ws9mMzWajsLCQDRs2sHnzZvz9/bnjjjt48cUXWbNmjVPPa6QMZwybBDzniHQyAK9IKd8VQuwD1gshHgR2Af92bP9v4AUhRAlqhHE1gJSyQAjxCrAPsAI/dkx7aTSaEdDabaaqtZcVU+OO+iw6NIBVWRNYlhHDtrIW8irb2VvVTlZyBIvTYogYZ+LhboaanrJYLNx5553s3r0bo9FIUVHRUfudcsopPPTQQ1RVVXHppZeSmZnJp59+Sm5uLkuWLAGUICUkJLjlPI7HCUVDSpkHLBiivZQhop+klH3AFcc41kPAQyM3U6PR9FNQ04FBCGZNjDjmNlEhAZwzewJL02PZXt5CfnUH+dUdzEqKYEl6DJHBJ7d4nGhE4E4effRREhMT2bNnD3a7naCgoKO2ueaaa1i6dCnvvfce5557Lv/617+QUnLDDTfw8MMPe8DqY6NzT2k0PoTdLtlX205aXAhhw3B2R4b4c9asRH6wIo2s5Aj21Xbw7OZyNu2rp63H7AaLNe3t7SQlJWEwGHjhhReG9EuUlpaSkZHBXXfdxYUXXkheXh4rV65k48aNNDQ0ANDS0kJFRYW7zT8KLRoajQ9R1txNt8lGVnLkiPaLCPLnuzMSuXFFGnNTI9lf28FzWyr4qKCO1m4tHq7kjjvu4LnnnmPZsmUUFRURGhp61DYbNmwgKyuL+fPns3//ftasWcOsWbN48MEHOeecc5g7dy5nn302tbW1HjiDI9FFmHwEXWDmMOP5Wry1u5qGDhM3nZqOwSBGfS26TFZyK1rZW9WG1S6ZMSGc7PRYYkIDXGC1e/jwww9ZtWqVp81wO4WFhcyceeQqCFcWYdLB3BqNj9DZZ6GsqZslaTEYDGNLHxEW6Md3psWzeHI0uRWt5FW1sb+uk2mJ4WSnxxAXFugkqzUnG1o0NBofobC2Eylh9nEc4CMlNNCP06fFszgtmp0VbeypauNAXSeZiWEsTY8lPlyLh+ZItGhoND6AlJL86nZSY0KICnH+FFJIgB+nZsaxaHI0uw61squyjeL6LqYkhLEsPYaEiKMjfjTjEy0aGo0PUNXaS3uvheVTY13aT3CAkeVT41g4OZpdh9rYVdnKiw1dZMSHsjQ9lgmRWjzGO1o0NBofIL+6nSB/I1Pjw9zSX5C/kVOmxLJgUhR7KtvYeaiNl7cdIj0ulKUZMSRFBrvFDo33oUVDo/Fyes02ihu6mJMSiZ/RvVHyQf5GlmbEMn9SFHlV7eRWtLJ+WyWTY0NYmhFLcpQWj/GGFg2NxsvZX9eBzS7JmjiytRnOJNDPyJK0GOalRJFX1UZuRSuvbK8kNSaEpekxpMaEeMw2b6G5uZmVK1cCUFdXh9FoJD4+HoBt27YREOC74cyD0aKh0XgxUkryazqYEBnkFZFMAX4GFqfFMC+1f+TRwsbcKpKjg1mWHktqTPC4q57XT2xs7EDeqfvvv5+wsDDuueeeI7aRUiKlxGDw3XXVvmu5RjMOqO8w0dRp8ugoYyj8jQYWTY7mxhXpnDE9nvYeC6/trOLVHVVUNHeP28WXQ1FSUkJWVha33XYbCxcupLKykqioqIHP169fz8033wxAfX09l156KYsXLyY7O5utW7d6yuxjokcaGo0Xk1/djr9RMG2CexzgI8XfaGDBpGjmJEdSUNPB9vIWXt9ZTVJkEEszYkmLDfHMyOOD+6Bur3OPOWEOnPfHUe26b98+nnnmGZ544gms1mPXObnrrru49957WbZsGeXl5VxwwQXk5+eP1mKXoEVDo/FSzFY7B+rVKu1AP6OnzTkufkYD81KjmD0xgsLaTraVt/DmrmoSI4JYmhFDRlzouJ22ApgyZcpAivPj8cknn3DgwIGB962trfT29hIc7D0BB1o0NBovpai+E7PVPuLkhJ7Ez2hgTkoksyZGUFjbwbayFt7eXUN8eCDLMmKYEh/mHvEY5YjAVQxOUmgwGI6Yvuvr6xv4XUrp9U5z7dPQaLyUgpp2YkIDSPLBBXVGgyArOZIblqdxzuxELDY77+ypZV3OIYrqO8e1z8NgMBAdHU1xcTF2u5033nhj4LOzzjqLtWvXDrz3xnrjWjQ0Gi+kuctETVsfWckRPj2tYzQIZk+M5IZT0liVNQG7XfJeXi0vbK1gf10Hdvv4FI9HHnmEVatWsXLlSlJSUgba165dy+bNm5k7dy6zZs3iqaee8qCVQ6NTo/sI4zkd+LcZD9fii6JG9lS2cfNp6YQEHHsW2deuhd0uKW7oYltZM01dZmJCA1iSFsOMCeFjztyrU6MfRqdG12jGEVabncLaDjLiQ48rGL6IwSCYPiGcaYlhlDR0sbWshY8K6sgpayY7PYYZEyIwjlE8NK7l5PpGajQnAaVN3fSabV63NsOZCCHITAxnakIYBxu7ySlr5uOCenJKW8hOj2FmkhYPb0WLhkbjZRTUtBMe5MekcZCaQwjB1IQwpsSHUtrUTU5pC5v21bO1VI08ZiVFuD3flub4aNHQaLyI9l4LFc09LE2PHfMcvy8hhGBKfBgZcaGUN/eQU9rMp4UNbCtrYXFaDFkTtXh4C1o0NBovYl9NBwCzk51Xnc+XEEKQHhdKWmwIh1p6yClt4bP9DWwva2FRmlp57q/Fw6No0dBovAS7XVJQ087k2BAigvw9bY5HEUIwOTaUSTEhVLX2srW0mS8ONLKjvIVFk6OZkxxFgJ8WD09wwqsuhEgVQnwmhCgUQhQIIe52tN8vhKgWQux2vM4ftM+vhRAlQogDQohzB7WvcrSVCCHuc80paTS+yaGWHjr7rCe1A3ykCCFIjQnhisWpXL4ohdjQQL4sauKZzWXsKG/BbLV72sQjMBqNzJ8/f+BVXl5+zG3Ly8vJyspyn3FOYjgjDSvwCynlTiFEOJArhNjk+OxRKeWfB28shJgFXA3MBiYCnwghpjk+XgucDVQB24UQb0sp9znjRDQaXye/pp3gACMZbqrO52ukxoSQGhNCdVsvOaXNfFXcxI6KVhZOimZeqncIbXBwsFeu4nYmJxxpSClrpZQ7Hb93AoVA8nF2uQhYL6U0SSnLgBIg2/EqkVKWSinNwHrHthrNuKfHbOVgQzezdKjpCUmOCubShSlcnZ3KhIggNpc08fTX5QBeucK8vLyc0047jYULF7Jw4UK2bNly1DYFBQVkZ2czf/585s6dS3FxMQDr1q0baL/11lux2Wwn7G9/XYfTz2EwI/JpCCHSgAVADrACuFMIsQbYgRqNtKIEZXAS+CoOi0zlt9qXDtHHLcAtI7FLo/F1Cms7sEvJ7Inj0wE+GpIig7l4QTL1HX1sLW2GOmjqNhHib2R7eStNXSan9hcfHsgZ0xOOu01vby/z588HID09nTfeeIOEhAQ2bdpEUFAQxcXFrF69mh07dhyx3xNPPMHdd9/Ntddei9lsxmazUVhYyIYNG9i8eTP+/v7ccccdvPjii6xZs2bIvqWUvLytkgfeKXDOCR+DYYuGECIMeA34qZSyQwjxOPC/gHT8/AvwQ2CoxyTJ0KOaox4LpJRPAk86+vS+xwaNxslIKcmv7iA5KpjYMM9X5/M1EiOCuGh+Mh9+uJcAo4Fus42OXgsmi40APyPuTN011PSUxWLhzjvvZPfu3RiNRoqKio7a75RTTuGhhx6iqqqKSy+9lMzMTD799FNyc3MHUqr39vaSkDC0aHX0Wfj163t5L6+W0zLjOLoH5zEs0RBC+KME40Up5esAUsr6QZ8/BbzreFsFpA7aPQWocfx+rHaNZtxS095HS7eZc2YnetoUnycqJACLzc53psdjstoRQHCAkZAAP49N+z366KMkJiayZ88e7HY7QUFHZy2+5pprWLp0Ke+99x7nnnsu//rXv5BScsMNN/Dwww8f9/hmq50LHvua6rZe7l01ndtOn8K6m111NsOLnhLAv4FCKeVfB7UnDdrsEqC/vNTbwNVCiEAhRDqQCWwDtgOZQoh0IUQAyln+tnNOQ6PxXfKr2wnwM5CZEO5pU04K/I0GokICiAkNINDPQI/ZRnOXic4+CzYP+Dza29tJSkrCYDDwwgsvDOmXKC0tJSMjg7vuuosLL7yQvLw8Vq5cycaNG2loaACgpaWFioqKgX2klDR2mmjsMmG12dlwyzLuOGOqyxeFDmeksQK4HtgrhOgfd/0GWC2EmI+aYioHbgWQUhYIIV4B9qEir34spbQBCCHuBD4CjMDTUkrXTr5pNF5On8VGcX0nM5Mi9LoDJ+NvNBAZEkCozU632UaP4xXi5pHHHXfcwWWXXcarr77KmWeeeURBpn42bNjAunXr8Pf3Z8KECfz2t78lJiaGBx98kHPOOQe73Y6/vz9r165l8uTJWG12qlp76eizEORn4P27TyMqxD2Fm3RqdB/B11Jgu5KT6VrkVbXxaWED1yydRGLEyIstnUzXYqycKDW61SEefRb1pB/sbyQ00IjR4Fti3W2ycqilB6tdkhQZRGNlqU6NrtGMF/KrO4gLDyQhXDvAXY2f0UBksIHQACM9Zhu9FvUK9jcSEmjEz8vFQ0pJQ6eJho4+AvwMTI0PJTjAj0Y326FFQ6PxEA2dfdR39HHG9Hifrs7na/gZDUQEGwgJNNJj+pZ4BBi9MjGixWansqWHLpOVqJAAkqOCPebY16Kh0XiIguoO/AyCmUl6bYYn8DMo8QgNNNI9SDyC/I2EepF4dPZZqGzpxS4lKdEhRIf4e/QhQ4uGRuMBLDY7hXUdTE0II8jf6GlzxjXGQeLR7yzv8wLxsEtJfUcfjZ0mgvyNTIoJ9YrvihYNjcYDlDR0YbLYyUr2jpxJGiUe4UEGQgL86DFb6e0XDz8DoYF+bhUPs9XGoZZeesxWYkIDmBgZ7DX1VbRoaDQeoKCmg6gQf1Kigz1tiuZbGA2C8CD/I8XDaibQIR6urufR3muhqrUHJEyKCXFbKO1w0aKh0biZth4zlS09rJgapx3gXky/eIQ6xKPHbMN0HPFobm5m5cqVANTV1WE0GomPjwdg27ZtBAQc/+Zvl5K69j6aukwE+xuZFBNCoBdMR30bLRoajZspqOnAIASzdHJCn8BgEIQNGnn0WGyYuo8Wj9jY2IG8U/fffz9hYWHcc889RxxLSomUEsO3wntNFhuHWnrotdiICwtkQmQQBi99oPCO8ACNZpzQX50vLS6EsED9zOZL9ItHXGggoYF+mG12WrrNtPaYj1sMqqSkhKysLG677TYWLlxIZWUlUVFRA5//+7l1XLPmRsw2OyHWLu784bVkL1lCdnY2W7duPeZxPYX+1mo0bqSsuZtuk007wF1N8SfQVX/i7UZCWCJknqXEI9CPkAAjvWYb3WYrrT1mAvwMhAb4DZkOZt++fTzzzDM88cQTWK1WAGx2SW1bL01dJowGQWZCONdfewv33nsvy5Yto7y8nAsuuID8/PyjjudJtGhoNG4kv7qdsEA/0mOPzj+k8S0MQhAa6EewQzx6+sXDaMBmP3LkMWXKlIEU5/2UNHRhstqIDPYnPEiJzSeffMKBAwcGtmltbaW3t5fgYO8JmNCiodG4ic4+C2VN3SxJi/Ga8MmTlsyz3NbVUOLRa7Hjb7ZittqQUg4kKZRS0tqrsu3a7JL0uFB2GuwDARFSymE5zT2J9mloNG6isLYTKdHV+U5S+sUjNiyQAKMBu4TWHgvtvRakBJtdpQKpbTcRGRUFHbWEBhh54403Bo5x1llnsXbt2oH33lhvXIuGRuMGVHW+dlK9MO5e41wMQjj8G0bCg/yw2cFqt3Ogrov2XgsTIoL48/89wve/dz4rV64kJSVlYN+1a9eyefNm5s6dy6xZs3jqqac8eCZDo1Oj+wg6BfZhfPFaVLb0sDG3ivPmTGDGBOeNNHzxWriKE6VG9wRSSpq6TNS192EwCCKC/AkJNBLmcJg7Y51OYWGhTo2u0Zxs5Fe3E+RvZGp8mKdN0biJwYWSIoLU6n+LzU63yUZbrwU/g5rOCnSSeLgLLRoajYvps9goaegiKyXSazKnalzL4EJJE6OCiQ0NQAiBn9FAkL+RPouNbrONdh8UDy0aGo2LKaztwGqXZE3UazNcid1uR0rp0Rtvf93u+g4TAX5ioFDSYIQQBAf4OcTDTrfZOmrx8MTUpBYNjcaFSCnJr+kgMSKIeF2dz6WUlJTQ3NxMbGysR4TjiEJJwQEkRwcdt5SsEg8jQf4G+qx2uk1KPIwGQWiAH0H+xxcPKSXNzc0EBY28TPBY0I5wH0E7PA/jS9eirr2Pl7cdYuXMBOamRJ14hxHiS9fC1cTExJCbm0tfX5/b++6z2GjtMSMlRAb7EzrKFDEWmx2T1Y7dLjEYBIF+BvwMBo6lHUFBQaSkpODv739Eu3aEazQ+Sn51O/5GwfQJ4Z425aSntbWV9PR0t/Zptdn566YiHv+igsyEMP5xzUKmJY7tby2l5GBjF1tLW2jsNBEZ7E92egwzkyI8VuJ1MFo0NBoXYbbaOVDfSWZiOIF+3pfiWjM2qtt6uevlXeRWtLI6O5XfXjCb4ICx/52FEExNCGdKfBilTd3klLawaV89OWUtZKfFMGuiZ8VDi4ZG4yKK6jsxW3V1vpORjwvq+OXGPGx2yWOrF3DhvIlO70MIwZT4MDLiQilv7mFraTOfFNaTU9bMkrQYZk+M8Eg0nhYNjcZFFNS0O0p1utdRqXEdJquNh9/fz7NbypmTHMnfVy8gLc61ySeFEKTHhZIWG8KhFiUe/9nfwPbyFhZNjiYrOdLl1QQHc8KehBCpQojPhBCFQogCIcTdjvYYIcQmIUSx42e0o10IIR4TQpQIIfKEEAsHHesGx/bFQogbXHdaGo1nae4yUdPWR1ZyhE/E3mtOTFlTN5c9voVnt5TzwxXpbLz9FJcLxmCEEEyODeXKxalctjCFiGB/Pj/QyDOby8itaMViO3ZND6facaLICyFEEpAkpdwphAgHcoGLgR8ALVLKPwoh7gOipZS/EkKcD/wEOB9YCvxNSrlUCBED7AAWA9JxnEVSytbj9K2jpxzoKJnD+MK1+LKokd2Vbdx8WjohAa4b0PvCtXAXrrwWb+2u5jev78XPaODPV8zj7FmJLulnpFS29JBT1kJlSw8hAUYWTY5mbkoUgf5Gz0VPSSlrgVrH751CiEIgGbgIOMOx2XPA58CvHO3PO+72W4UQUQ7hOQPYJKVsARBCbAJWAS878Xw0Go9js0v21XaQER/qUsGoa1ehpTmlzcxLjSLIC+tJ+zq9Zhv3v13Ahh2VLJ4czWOrFzAxyntqW6TGhJAaE0J1Wy85pc18VdzEjoIDJ95xDIzoGy2ESAMWADlAokNQkFLWCiESHJslA5WDdqtytB2r/dt93ALcMhK7NBpvorSxi16zzaUrwHvNNn747HYArnpyKwF+BhakRrE0I5Zl6TEsnBytRWSMFNV38uMXd1LS2MWdZ07lp2dlem0amOSoYC5dmEJNbS3bnnvIpX0NWzSEEGHAa8BPpZQdx5mnHeoDeZz2IxukfBJ40tGnHndrfI78mnbCg/yYFBPikuNLKfnlxj0U1nUA8NSaxeSUNpNT1sI//lPMYxL8jYJ5KVEszYhhWUYsiyZHu3TUczIhpWT99koeeKeAsEB/XvjhUk7NjPO0WSfGZmXiptu52Py1S7sZ1rdICOGPEowXpZSvO5rrhRBJjlFGEtDgaK8CUgftngLUONrP+Fb756M3XaPxPtp7LVQ097A0PdZl1fke/+Ig7+bV8qtVM7jjj3D2rMSBOfaOPgu55a1sLWtma2kLT3xRytrPDuJnEMxJiWRpeixLM2JYPDma8CD/E/Q0/ujss/CbN/J5Z08Np2XG8Zcr55EQ7iPRb5/8Dko/gwv/Ab9b47JuhuMIFyifRYuU8qeD2v8ENA9yhMdIKe8VQnwPuJPDjvDHpJTZDkd4LtAfTbUT5QhvOU7f2hHuQDs8D+PN1+Kbg83klDVz44p0IoOdf1P+z/56bnpuBxfMnchjV8/HYDAc91p0mazkVrQOjETyqtqw2CQGAVnJkSxNj2FpeixL0mNcYq87Gev3Ym9VO3e+vJOq1l5+fvY0bv/OFN8py7tnPbxxK2TfCuf/n0vTiAxHNE4FvgL2Av0xXb9B+TVeASYBh4ArpJQtDpH5B8rJ3QPcKKXc4TjWDx37AjwkpXzmBH1r0XDgzTdKd+Ot18Julzy9uYzYsAAuWZBy4h1GSElDF5es3cyk2BA23rac4ADjiK9Fr9nGzkNKRLaWtbD7UBtmmx0hYOaECJZmKBFZmh5DdKhvVRgc7fdCSskzm8t5+INC4sMCeWz1AhanxbjAQhdRnQtPnwep2XD9G2D096xoeBItGofx1hulJ/DWa1He1M0bu6q5YG4SmWPMP/Rt2nstXLJ2M+29Ft7+yakkOyJ4xnot+iw2dle2kVPawtbSZnYeasVkVc+GMyaEq5FIRizZ6THEhXl3lt7RXIvWbjO/3JjHJ4X1nDUzkT9fMde3yvF21sOTZ4DBD275HEJjAdcmLNSi4SN4643SE3jrtXg3r4aq1l5+dFqGU3MD2eySm5/bzlfFTbz0o2Vkpx9+Cnb2tTBZbeRVtQ9MZ+0ob6XXYgNgakLYgIgsS48hIcK75vpHei22l7dw18u7aO4y8+vzZ/CD5Wm+tRDTaoLnvg91e+Gmj2HCnIGPdJZbjcbL6TFbKW3sZn5qlNOTyf354wN8dqCRBy/OOkIwXEGgn5ElaTEsSYvhTlSq7r3V7eSUtpBT1sxbu2t4MecQAOlxoSzrn87KiCEp0nvWLxwPu13y+BcH+eumIlKig3nt9uXMSfGx/GBSwnu/gMocuOLZIwTD1WjR0GicQGFtBza7ZPbECKce963d1Tz++UGuWTqJ65ZNduqxh4O/0cDCSdEsnBTN7WdMwWqzs6+2g62lzeSUtvBuXi0vb1PLrybFhAyMRJamx5DqopDjsdDQ2cfPN+zh65ImLpw3kYcuyfLNKLLt/4JdL8Dpv4TZl7i1az095SN465SMJ/C2ayGl5PlvKgjyN3DVkklOO25+dTuXP7GFOcmRvHjzMgL8jl5Y5ulrYbNLCms7yClrIae0mW3lLbT1WAC14EyJiBqNTI4Ncen0z4muxVfFjfxsw266TFYeuHA2Vy5O9a3pqH7KvoLnL4LMc+Dql2CI6oDap6Hx+M3Bm/C2a1Hd1ssr2ys5e1ai09KgN3WZuOgfm7FLydt3nnrMUrHedi3sdklRQ+fAdFZOaQvN3WYAJkQEHY7OyoghIy7UqTftY10Lq83Oo58U8c/PDzI1Poy11469UJLHaK1Qju/QeLj5EwgaemSrfRoajReTX91OgJ/BaTcis9XOHet20tRlYuNty32qtrjBIJgxIYIZEyK4YXnaQBW6b0rVSGTLQeUXAYgPDyQ7PYZljimtzIQwpz/51zgKJe2oaOXqJan87vvOKZTkEczdsP5akDZY/fIxBcPVaNHQaMZAn8VGcX0nMyZEDDl9NBp+/24B28pb+NvV833PQfst+qvQTU0I5/plk5FSUtbUPTCdlVPWwnt5tQDEhAaQnXb3JTVhAAAgAElEQVR4OmvGhPAxLa7btK+ee17d49JCSW5DSnjzDmgogGtfhdgpHjNFi4ZGMwaK6jux2KTTpqVeyjnEuq2HuPU7GVw0/6h8nj6PEIKM+DAy4sNYnT0JKSWVLb1sdUxl5ZQ182FBHQBRIf4sSYthabrKnzXcGtkmq40/frCfZzaXk5UcwT9WL3Rr3QuX8NVfYN+bcPbvYepZHjVF+zR8BG+bu/Yk3nQtXso5hE1Krls6acxTK9vLW7jmqa0snxLH0z9YMqwbpDddC2dR1dpz2CdS1kJFcw8A4UF+AyKyNCOWrG+VOxVCUNbYxZ0v7yS/uoMbV6Rx33kzfL8++4EP4eWrYc7lcOlTMIzvmfZpaDReSENnH/UdfZwxPX7MglHT1svt63JJiQ7hsasXOH2thy+REh1CyqIQLlukUrHUtfeR40jAmFOmSp0ChAYYWTQwElHrVy74+9cYDYKn1iz2mkJJY6LxALx2MyTNhQv/PizBcDVaNDSaUVJQ04GfQTAzaWwOyV6zjVte2EGfxc76WxYRGeKD6wZcyITIIC6anzwwXdfQ2ce2spaB1Cd/+uhw0aEZE8L52+oFA2lWfJreNnh5NfgHqdBaf+84Jz095SOcjNMQo8UbroXFZuepr0pJjw3lvDlJoz6OlJKfbtjN23tqeOr6xZw1wqdjb7gWnqa5y8S2shbOnzsRi9XmtYWSRoTdBi9dCaVfwA3vwORTRrS7np7SaLyMg41dmCz2MTvAn/yylLd21/DLc6ePWDA0itiwwAHhPikEA+DTB6DkE7jg/xuxYLiak+QKazTuJb+6g8hgf1KiRz9l8PmBBh75cD/fm5PEHWd4LoRS42XkvQqb/waLb4LFN3ramqPQoqHRjJC2HjOVLT1kJUeO2gFe2tjFT17exbTEcP50xVzfTGehcT41u+DtO2HyClj1R09bMyRaNDSaEVJQ04EQMGuUyQk7+yz86Pkd+DmifHTtbg0AXQ1qxXdoPFzxHPh5Z10P/W3VaEaA3S4pqGknPS6UsMCR//vY7ZKfrt9NeXMP625a6pWZYDUewGqGV9ZATwvc9BGExXvaomOiRUOjGQFlzd10m2yjdoD/dVMRn+5v4IELZ3PKlFgnW6fxWT64Fw59A5f9G5Lmedqa46KnpzSaEZBf3U5YoB/psSNPS/FeXi3/+KyEqxansuYU99fG0Hgp2/8Nuc/AqT9Tq769HC0aGs0w6TJZKW/qYdbEiBEn0ttX08E9r+5h4aQofn/xbO341igqtqhRRuY58N3/8bQ1w0KLhkYzTPbVdGCXI6/O19Jt5kfP7yAy2J8nrlvk+7mQNM6hrRI2XA/R6XDZv8DgG98L7dPQaIaBlMoBnhIdTFTI8KNaLDY7d7yYS2OXiVdvPYWEiCAXWqnxGcw9sP4asJkdtTF8JwW+HmloNMOgqrWXth7LiB3gD71XyNbSFv546RzmpUa5yDqNTyElvP0TqNurRhhxmZ62aERo0dBohkF+dTuB/gamJoQNe58N2w/x7JZybj41nUsXprjQOo1PsflvkL8RVv4Wpp3raWtGzAlFQwjxtBCiQQiRP6jtfiFEtRBit+N1/qDPfi2EKBFCHBBCnDuofZWjrUQIcZ/zT0WjcQ19FhslDV3MnBCB/zBzG+VWtPDfb+ZzWmYc9503w8UWanyGoo/hk/th9qUqWsoHGc5/wLPAqiHaH5VSzne83gcQQswCrgZmO/b5pxDCKIQwAmuB84BZwGrHthqN11NY24HVLpmdPDwHeG17L7e+sJOJUcH8ffWCkyeJnmZsNBWr2hgTsuCitV5RG2M0nPDbLKX8EmgZ5vEuAtZLKU1SyjKgBMh2vEqklKVSSjOw3rGtRuPVSCnJr+kgMSKIhPATO7H7LDZueyGXXrOVp9YsHpHTfMSYOvuNdF0fGufQ165qYxj9VG2MAN/NBDCWR6A7hRB5jumraEdbMlA5aJsqR9ux2o9CCHGLEGKHEGLHGGzTaJxCfYeJpk4TWcMYZUgp+c3re9lT1c5fr5rPtMRw1xlm7obdL6nfSz93XT+asWO3wWs/gtYyuPJ5iJrkaYvGxGhF43FgCjAfqAX+4mgfarwlj9N+dKOUT0opF7uqgIhGMxIKatrxNwqmTzixAPz76zJe31XNz86axrmzJ7jOKKsJ8jaAqUO9P7QVavNc159mbHz2EBR/BOc9AmmnetqaMTMq0ZBS1kspbVJKO/AUavoJ1AgiddCmKUDNcdo1Gq/FbLWzv66TzMTwEy7I+6q4kT+8X8iq2RP4yXenus4omxX2boSuRuVMBYhOg6IPoe2Q6/rVjI781+Crv8CiH6j6GCcBoxINIcTg+paXAP2RVW8DVwshAoUQ6UAmsA3YDmQKIdKFEAEoZ/nbozdbo3E9xQ2dmK0nrs5X0dzNnS/tIjMhnL9cOW/EKUaGjd0O+95U4jDjexDrKNw0+2IIilI3qJ7huh81Lqc2D978MaQug/P+5LOO728znJDbl4FvgOlCiCohxE3A/wkh9goh8oAzgZ8BSCkLgFeAfcCHwI8dIxIrcCfwEVAIvOLYVqPxWgqqO4gJDWBi5LEd4F0mKz96fgdCwFNrFhM6inTpw0JKKPpAReBknq0icPrxD3YkuhNqFGLpdY0NmuHT3aRqY4TEwFUveG1tjNEgvLkovRBCerN97kQIgb4WCndci+YuE89/U8Hp0+JYNDlmyG3sdslt63L5dH8Dz/8wmxVT41xn0MH/wKEcSFsB6acPNB9xLdoOwZ71EJkCc6/ymVxGzsJr/kdsFnj+YqjeATd+AMkL3W6CECLXVX5hHUCu0QxBQU0HBiGYmXTsqKm/fVrMx/vq+a/zZ7pWMA5tVYKRvAjSTjv2dlGTYNoqaK2A4o91KK6n+PDXUPE1XPh3jwiGq9EJCzWab2GzS/bVdjAlIfSYpVg/zK/lb58Wc/miFG5ckeY6Y2p2w8HPIHGWmpY60bx40lzobYGKbyAkFlKzj7+9xrnkPgvbn4LlP4G5V3raGpegRUOj+RaljV30mm1kTRzaAb6/roOfv7KHealRPHhxlutqYzQeUFFRMRkw44LhO1LTv6Mc4gf/A8HRPpcQz2c5tBXeuwemfBfOesDT1rgMPT2l0XyL/Jp2woP8mDRE/e5WR22MsEA/nrx+EUH+LvIbtJbDvrcgYiLMvmRk/gkhYOb3ISxRHaOz3jU2ag7TXq1qY0SlwuVPn9T+JC0aGs0gOvosVDT3MHti5FGhs1abnTtf3kl9u4knrl9EoqtqY3TUqCio4GiYc8XoIm+M/iqiyi8I9r56OOWIxvlYemHDtern1S+rv9tJjBYNjWYQBdVqlfWsIarz/eH9/WwuaebBS7JYOMlFN4buZsh7BfxDYN7VKpx2tASGK9Gx9qk1HDaL8+zUKKSEd+5WvqfLnoKEkz+jsRYNjcaB3a6q802KCSEy2P+IzzbmVvH05jJ+sDyNKxenHuMIY6SvHfa8rKaX5l2tbvpjJTwRZl0EnXVQ+I6OqHI23/xDpXQ5879g+nmetsYtaNHQaBwcaumhs8961ArwXYda+c0be1k+JZb/+t5M13Ru7oY9G8BmgrlXq0VhziIuEzLOVI71si+dd9zxTsmnsOm3SpRPv8fT1rgNHT2l0TjIr2knOMBIRlzoQFt9Rx+3vpBLYkQga69ZOOwiTCPCalJTUn3tMO8qNTpwNqnZ0NMMFVuUIE2Y4/w+xhPNB2HjjZAwCy7650mTImQ46JGGRgP0mK2UNnYzMylioGhSn8XGrS/k0mVStTGiQ12QCsJmVf6GrgYVJeWqtNlCqNKi0ZPhwAc6ueFY6OtQtTGEEa5+EQKHXwLYLbh4ClKLhkaDqs5ns0uyHA5wKSX//WY+uyvb+MsV85gxYXhV+0aE3Q6Fb6kV3DPOhzgXZscFFQY6+xIIioT813Vyw9Fgt8Mbt0JzCVz5nMow7E1IqbIBuBAtGppxj5SS/OoOJkYFERsWCMCzW8rZmFvFXd+dynlzkk5whFF1qhbuNRbB1LPcN13kH6wiqpCO5IZ97un3ZOGLP8KB92HVw0fkAPMK7DYV7FC906XdaNHQjHtq2vto6TYz27ECfEtJEw++V8hZMxP56VnTXNNp6edQuwcmL4fUJa7p41iExKhaHH1tKtW63ebe/n2VfW/BF4/Agusg+xZPW3MkNisUvAH1BZDxHZd2pUVDM+4pqG4nwM/AtMRwKlt6uOOlnWTEhfLoVS6qjXEoR6WcmLjAc0+r0ZOVj6OlDIo36VDcE1GXD2/cDilL4Ht/9S7Ht9UEe19RafOnnaMeRFyIjp7SjGtMVhtF9Z3MmBCBxWbnR8/vwG6XPLVmMeFB/ic+wEip3aNyQiXMgMxzPHvzSZqn/BqHtjqSG7p5xOMrdDfD+tUQFAFXrQO/QE9bdBhzjxKMznqVOmZwnRUXoUVDM64pquvCYpPMnhjBPa/uoai+k2dvzCZtUNit02gsUpFLMekw80IweMFAP+MMlRX34KeO5IYudsb7GjYrbPyBuinf+AGEu7D2+0gxdar6Kb1tkHWp2xJTesG3VqPxHPk17cSFB/Lqjko+yK/j1+fN5PRp8c7vqD8BYXiS8id4S0I7IWDG9yEsQfk3uho8bZF38fF/qwWR3/8bpCzytDWH6W2FXevU2p65V7g1k7EWDc24paGzj7r2Prr7rPz1k2IuWZDMzaelO7+jjlq1FiM4avQJCF2JXwBkXa6mXfa+CqYuT1vkHexaBzmPw7Ifw/zVnrbmMF2NyjZrH8y/xu1hv1o0NOOWgpoO2nrNPPZpEXNTInn40jnOr43R3axyE/kFqxKsAUenW/cKgiKUoFl6dHJDgMrt8O7P1PTd2b/3tDWH6aiB3evU7/OvU6nz3YwWDc24xGKzs6uilffyagkJ9Of/uaI2Rl8H5K0/nIAwyAULBJ1J+ASYeRF01sL+d8dvRFVHLWxw3JAvfwaMXuL6bS2H3S+pdPcLroMwF0yjDgMtGppxSVFdJ2/urqatx8wT1y0kKXIMKciHwtyjRhjWPjXCcGYCQlcSP009XTfsh/KvPG2N+7H0KcEwdaraGN7yd2sqhrxX1Wr+Bdd5tGaHl0ioRuNe/vTRAQ619PKHS7JYnObkG4PVrMIge9tUnWhvirgZDqlLVXLD8s0QHOOWME6vQEo1JVW9Q4XWJs7ytEWKunzY/55KZDnnSo9PcWrR0Iw71m0t5/OiRs6dncg1Syc79+D9CQg761UYZLSTj+8OhIBpq5ToHXhfPd1GuaiGiDeR8wTseQnO+LVa8+ANVOeqxZdRkyDrMq9YI6KnpzTjiryqNu5/ex/JUcH88bK5zj243Q6Fb6u55+nnuTUM0ukYjEr0giKVCPa2etoi13LwM/jov2DGBXD6vZ62Ro16KrZA0ccQO1WNMLxAMGAYoiGEeFoI0SCEyB/UFiOE2CSEKHb8jHa0CyHEY0KIEiFEnhBi4aB9bnBsXyyEuME1p6PRHJuGzj5+9NwOQgKM3H7GFKJDnBj6KiUUf6QKHU1dCUlOFiRPMF6SG7aUqdoYcdPgkic8v+hSSij9DEq/gMTZKjOxtzjjGd5I41lg1bfa7gM+lVJmAp863gOcB2Q6XrcAj4MSGeB3wFIgG/hdv9BoNO7AZLVx+7qdtPaauWBuEssyYp3bQdkXqk705FNUwaOThZAYddPqaXEkN7R72iLnYuqC9deoG/Xql5xTYncs2O0q+/GhHEheqKbJvGUhqIMTioaU8kvg24n3LwKec/z+HHDxoPbnpWIrECWESALOBTZJKVuklK3AJo4WIo3GJUgp+d1bBeRWtHJN9iTS4kJJd2aakMptUPENTJwP6a7NMOoRotMOJzcs+cTT1jiP/toYjfvhimchJsPD9tjU9Gb/w4enc5Mdg9GOeRKllLUAUspaIUSCoz0ZqBy0XZWj7VjtRyGEuAU1StFonMK6rRWs317Jj05LJyzQn1lJkRidlb22Nk/Vio6fDpnneuU/uVOYOF/lqDqUo0YfKYs9bdHY+fJPaj3KuQ/DlDM9a4vNolKbNx9Utkxa5ll7joOzJ++G+o+Rx2k/ulHKJ6WUi6WUJ8G3UuNpvjnYzAPv7OO7MxJYOSMRu1TJCZ1CU7FKQBid5j0JCF1J+hnKuV/yibq5+TKF78Lnf4B518Cy2z1ri6VPrelpKYXpq7xaMGD0olHvmHbC8bM/y1kVMDg2LwWoOU67RuMyqlp7+PFLO5kUG8KjV82jsK6DlOhg59T6bq2AgjdV7HzWZV7lqHQZBoMSx4Hkho2etmh0NBSqaamJC+GCRz07OjR3qzDf9mp1bScu8Jwtw2S0ovE20B8BdQPw1qD2NY4oqmVAu2Ma6yPgHCFEtMMBfo6jTaNxCT1mK7c8n4vFZuepNYvp6LXS1mMhKzly7AfvrIP8jY4EhFd6XwJCV9Kf3NAY4JvJDXta4OXVEBAKV78I/kGes6WvA3a9qPKTzbncexYTnoDhhNy+DHwDTBdCVAkhbgL+CJwthCgGzna8B3gfKAVKgKeAOwCklC3A/wLbHa/fO9o0GqcjpeSXG/MorOvgsdULmBIfRn51O4H+BqYmhI3t4D0tjgSEQd6dgNCVBEUo4bB0O5IbWj1t0fCwWWHjD6GjWq349kCyvwF6WlSmWnMnzLsKYqd4zpYRcsIxtZTyWDmBVw6xrQR+fIzjPA08PSLrNJpR8M/PD/JeXi33nTeDM6cn0GexUdLQRVZyJP7GMfgd+jpU0RspYa4PJCB0JRFJajol/3U48J763duDAD75nVr/cOE/PBsW3dXg+B7ZYf61Ppdm5iT33GnGG58W1vPnjw9w4byJ3Hq6CqHcX9eJ1S6ZnTyGm7ylV40wLD1qhBHq5HUevkj8dJXcsH4flH/taWuOz+6X4Zt/QPatsPB6z9nRXqVGGMIAC673OcEAnXtKcxJR0tDF3et3Myspgkcum4sQAikle6vbSYwIIiF8lPPXVjPkvaJSacy9Uj1laxSTljmSG36tQnETZ3vaoqOpzoV37oa00+DchzxnR0uZ8oUFhKtU+cFRnrNlDOiRhuakoL3Xwi3P7yDQz8CTaxYTHKBW0TZ0mmjqNJE12lGG3QYFr6saE7MucnuVNK9HCJVnKyoV9r+vnqS9ic46WH+tinK74jkw+nvGjsYDKnAgONqR2tw3BQO0aGhOAmx2yd3rd3GopYfHr1tEctTh2hj51e34GwXTEkeRHsJuh8J31BPi9PPUdIzmaAxGVfc8MNyR3LDN0xYprCbYcL2qo331y56bUqzNUwv3wicoH0bgGIMxPIwWDY3P86ePDvD5gUbuv3A22emHa2OYrXb213WSmRg+8qp8UkLxxyqmf8qZkDTPyVafZASEqOSGdpt6oraaPGuPlPDeL6BqG1z8uOdqglTtULUwoiar4Al/Jxf78gBaNDQ+zVu7q3nii4Ncs3QS1y07snZFcUMnZqt9dGszyr+Cml1qzt7LV+h6DaGxg5IbvuXZ5Ibb/wW7XoDTfwmzLz7x9s5GSuXnKd6kqiHOueKkWc+jRUPjs+RXt/Or1/JYkhbN/d8/2gFbUN1BTGgAEyNH6ACv3K6q1iXNU9FBmuETkw7TzlFpRg5+6hkbyr6CD34F086DM37j/v6lVOde9hVMmAOzvCu1+Vg5ec5EM65o6jJxy/M7iAkJ4J/XLiLA78jnn+YuE9VtvZw+LQ4xkvUDdXtVbqX4aap6nbevPfBGJi5Qo43KbapcbMoi9/XdWgGvrFGFiy590v35wOx2KPpA+TFSFsPUs06675AWDY1Pcse6nTR3m3nt9uXEhx9d0aygpgODEMxMGkHUVFOxigCKngwzLzr5ExC6kowzVYhyyScqUshdK57XXwPSBqtfdv/iS5tVpTZvPABpp6rXSSYYoKenND7KtvIW/u/yuUP6K2x2SWFtBxnxoYQEDPO5qO2QSkAYljB+EhC6kv7khqFx7kluKB1Jsxv2weVPuz8th9Ws1mA0HlCji/TTTkrBAC0aGh+ix2zlf95UVYdv/U4GF80fsiQLpY1d9Jhtw3eAd9ariJ+gSLV4z0tqMfs8fgEqEZ/BX11fc7dzj99Zp8rQvvszWOtIC3LWA+qm7U4sfZC3XtWGn3E+pC5xb/9uRj9OaXyC3IoWfvHKHsqbewC499wZx9w2v6ad8CA/JscMI5lgfwJCY6BKHBfgxIp+GiXEcy6H3S+qNRzzrhn9KK69Gio2q6ik8q+hxVHTIyDcEeG2A5b/xGmmDwtTl/r+9DTDrIsh4djfy5MFLRoar6bPYuPRT4p46stSkiKDeelHS1nxCMesvNfRZ6GiuYfs9BgMJ6rOZ+oclDjuGnWD0zifiIkw4/tqgdtIkhu2VjhEYjNUfK2e5EH9nSYth8U3wuQVMGGuEqLrhXunhPra1ffH1KGE0dPlYt2EFg2N15Jf3c7PX9lNUX0Xq7NT+c35MwkPOn4aiILqDgBmTzyBAFh61T+8pUcJRmics8zWDEXCDOj9DpR+ASGxykk8GClV5boBkdgM7Y4K0cHRShyyb4W0FZCYpVahe5LuZjUlZTXBvNUQmeJZe9yIFg2N12Gx2Vn7WQn/+E8JMaEBPPODJZw5I+GE+9ntkoKadibFhBAZfBxxsZrVHHtvq1p05cm6CuOJSaeoaZyyr5QQGAMOTzVVbIFORzHPkDglDsvvUj/jZ3pXJFtnnZqSAkdq80TP2uNmtGhovIri+k5+/soe9la3c9H8iTxw4WyiQoa3kvZQSw+dfVZOnxZ/7I3sNjVN0lGjVi/HpDvJcs1xkRIa9yu/RPHHsOXvqogTQNgEJQ6TV6gRSNw07408aquEva+oIlzzVqvMvuMMLRoar8Bml/z761L+/HERYYF+/PPahZw/Z2QpyPNr2gkOMJIRdwxntpSw/101DaITELoWux0aCg77Iyq2qFEGQHgSxGWqfEyn/VytvPdWkRhM80GV8TgwUgVNjFMfmBYNjccpb+rmnlf3sKOilbNnJfKHS+YMuWDvePSYrZQ2djMvNQq/oarzSanyANXvU6lBJs53iu0aB3abWk1f/rXyR1RsgT5HttuoSZB57uHRRHSaEpCdz0P1DkiY6f1hzg2FKuNxSKyqhTGOo+y0aGg8ht0ueTGngj+8vx8/o+CvV87jkgXJI0v74aCwthObXZI18RirgMu/VsV4UrN1AkJnYLNC7R41iijfDIe2gqldfRaTATO/r6aaJq9QtTa+TWicmh7Me0UlN8y63Lv8FoOp2Q1FH0JEsvKB+Y+ymNdJghYNjUeoaevl3o15fF3SxGmZcfzf5XNJihxd2mgplQN8YlQQsWFDPLFW7VCikTQXpnzXN6ZCvA2rWWX97ReJyhwwd6nPYjMh61KHSCwffmBBTDpkng1FH8HB/0CmmxflDYdDOcq2mAx1jp4q4uRFaNHQuBUpJa/trOaBtwuwSclDl2RxTfakUY0u+qlt76O5y8zZs4aIYqnLV9NScZkq66kWjOFhNamRWflmlSa+arsKTwYVzTTvajWKmLxibNFDyQvVAsuq7RASDcluTG54PKSEsi/VNFvCDLW2xNNhvl6CFg2N22jo7OM3r+fzSWE92Wkx/PmKeUyKHcaq7ROQX91OgJ/h6Op8zQcdBXAmqdW63jr94Q1YetWNu3+NRNV2sPYBQq2LWLjGIRLLnb+mZcp3Vfhz8ScqFNfTi+T6/V/VucpJP22V/u4MQouGxi28l1fLf7+5l26zjf/+3kxuXJF+zFXdI8FktVFU38mMCRFHpkdvq1SRLmHxarWuTkB4JOZuNcXULxLVuWAzgzCoGhCLb1KO60mnuD6s1GCAWReqokkFbyqB8tRiS7tdrVqvy1f+Lz2deRT6P0njUlq7zfz27QLe2VPD3JRI/nrlPKYmjKJe9zEoquvCYpNHJifsalCL9wIjYO5V3h+Z4w5MnWp+vsKxmK5mF9itIIwqkmzpbconMWmZZ0JJ/QKVkzn3OfW3W7jG/RFKNqvKyNtUDOmnq1GVFoyjGJNoCCHKgU7ABlillIuFEDHABiANKAeulFK2CjVp/TfgfKAH+IGUcudY+td4N//ZX8+vXttLa7eZX5w9jdvPmDJ0OOwYyK9pJy48kMQIhzD0tqr0IMYAJRjjNTSyt01FNPU7rmv3qDoTBn/lR+hfbZ26FAKdJ+Jjoj+54a4XIf91tXjOXSNEq0klVGytgMxz3Fs4ysdwxl/kTCll06D39wGfSin/KIS4z/H+V8B5QKbjtRR43PFTc5LR2Wfhf9/dxys7qpgxIZxnfrBkdHW6T0BDZx917X2cMT1eOdIHEhDaVD6p4Cin9+m19LQop21/Fti6vYBU4pm8WC2iSzsVUrIhYOx+JJcRMRFmXqCmqYo+gBkXuP5p39KrQn8761TfE+a4tj8fxxUyfhFwhuP354DPUaJxEfC8lFICW4UQUUKIJCllrQts0HiILSVN/HJjHrXtvdxxxhTuPiuTQD8nRp3U7FY/P/sDBY3h+HUEM1M0QpEZ6vNVbYMJs2Hbk87r09v553K1+hpUeouUJXDGfcpxnbIY/EcXyuwxEmYqESz7UpWLTVvhur76HzR6W9W6kfhpruvrJGGsoiGBj4UQEvh/UsongcR+IZBS1goh+jPNJQOVg/atcrQdIRpCiFuAW8Zol8bN9JptPPLhfp7dUk56XCiv3racRZOjnXNwu10trvpmrZpuAayf/4n9tu8xRdQTVLXtyO3Lv3ROv75CWAJkXQKTT1VTTyeDD2fyckdywy+VIz5hpvP76G2FPRvUepM5V+g8ZMNkrKKxQkpZ4xCGTUKI/cfZdqgxpjyqQQnPkwAOMdJ4ObkVLdzzah5lTd38YHkav1o1g+AAJ4wuTF2w+yXIeVzli4pMhXMehAfuouT2Cvr21pE1PwlqP1Cfz7rINTcXb+YBAWve9LQVzkcImH6+qllR+K7ydzgzG2cGqKAAABSeSURBVHF3E+x5WQUDzFsNkUNXgdQczZhEQ0pZ4/jZIIR4A8gG6vunnYQQSUCDY/MqYHA+gRSgZiz9azyLyWrj0U3FPPnlwYECScunOCFUsr1aTS/lPqNuGsmL4fL/UQusjH7AXeRXdxAZ5Edq43/Ueoxp544/wTjZMfqpVdg7n1dlXRfd4JzIro5aldpcGGD+dSosWzNsRi0aQohQwCCl7HT8fg7we+Bt4Abgj46fbzl2eRu4UwixHuUAb9f+DN8lv7qdX7yyhwP1nVy9JJX/+t6JCySdkOqdsPWfKnW5tKv8RafcqeLlv0VlczfnBecj6osg4ztqWkZz8hEQqqaOdjpCcRdcP7bpt9YKyN8I/iEqum4cpjYfK2MZaSQCbzjSP/gBL0kpPxRCbAdeEULcBBwCrnBs/z4q3LYEFXJ74xj61ngIi83OPz87yN//UzyiAknHxG6DA+/DN/+EQ1tUvefsW2HpLSob6jFI6dxJujgI6cvUAjTNyctAcsNXYd/bkHXZ6FZoN5WoB5LgKCUYQcdIbqk5LkIFM3knQgjpzfa5EyEEnr4WxfWd/OLVPeRVjbxA0lGYumD3i2pk0VoOkZNg2W3qSfI4/8x2u8RoNFD40n3MmL/cPSGZXow3fC/cRnUuFH0MqUtg6tHJDY97LeoLlG8kLMGxfseLw46dgBAiV0q52BXH1ivCvZhuk5Xy5m7KmlSFsye/POgRO+wStpe18HlRIwFGwSULkpmZFM4rOypPvPO36WtXi86qdqjcRpGrYNYK5Y+wGGBbI9B41G7CbiGyt5KorlIAoibNVo7ScSwY447kRSoUt3K7qmsxccHw9qveqaoFRqaqxYMnQ3SZB9Gi4UVIKWnsNFHapISivqMPKSEsUP2ZMuLC3G5TQ2cfz2wuo7ihm3kpUaw5ZfLx628fi+YSNQ11aKt6n5oN07+nss8eA2G3ENxVQVh7CSGdZRjsFmx+as1BQvaVOuvoeGTKShUqW/QxBEWdOEy24hso/Rxip8Lsi3Vqcyegp6c8jNlqp7K1h7JGJRRdJitCQGJEEOlxoWTEhRIfHojBYHDrNISUknU5h/jDe4X4GQX3f382ly4cYYEku01lmf1mLVRuVbmgFq6BpbeqzLNDYbOq8NnGQpUDyGZRi9Pip0P8DIiajDAax8+UzAkYV9NT/VhNKqLK1AkLb4DQWOBb10JKJRaHtkLiLDWNOY4eMlw5PaVFwwO091ooa+qmrKmLqpZerHZJgJ+BybEhpMeFkhYbSmjgkYNAd94catp6+dVreXxVrAokPXLZXCZGjWBVcV8H7FoHOU9AW4WqBb3sdlhw3dB5jmxWaC1TJTWbi1XBnwGhmA5RaUc4PsfljfIYjNtr0dumIqqMgY7khiGHr4WUqrBTzS41hTXt3HE3jalFw8ex2yW1HX2O0UQXTV1mAKJC/B2jiTCSo4P///bOPziu6rrjnyPtrna1km3ZEgZs459gg8EEMOAaknFIMsRA+NFCsKGd0AHcBlqaNKSZAknpNAk0IQ0wccCUMC2eTjGG8puYJi0hMGCMoYAB/wD8G+NgWZaF9WPf27enf9wn7VpI8traX7LOZ+aN33v73tN9x7v3e++5954zYKjwUlQOvRMk3XTe8Vx55kEkSGrdCq8uCVuBbTBhDvzR9TDj/M+28voUijg0TndJb0ZN7LdlOGwryj4Y1rbY+5Fb/DniKJi1AIlE0SAN6552ueCPmePywQ8zwQAbCB+SdPmBG8Te1c7m3R10+QFVIoxrSPCF40YypTFJQ/IQZx4VgdwESadPauCOy05m4pg8I8Ruew1WLnbTIcH5judc/9lIoUHazZTatRaaN2SFommG2xomDSsXgjFIRo5zDZL3nnBhZsBFx939gROLiTYVuxiYaBQIVaWl3WNTczsbm9vZ0dqJKiRi1a430ZTkmNG1xKOVVyk+u+Zjbn7sIBMkBWlY95RbX7F9FdSMdL2KMxbBqJyF/5nACcUn3UKRcrNXTCiMQjD2BOhsgU0vuuOWMDqALfYsGiYagyAdZNi+p7NHKNo6fQCa6ms4Y9JoJjclGVsfp6oAGeqKQWuHxw+eeJcnDyZBUleby7C28l7Yu9VV+vN/4kKRd49XdAvFrnWwa31WKBqPc1NrTSiMQjLxLDcVF1wUgbEzy1uewxwb0zhI9qXSbA5FYuvudvxAiVYLE0aHg9iNSUYMNpxGHxTad52bIOmGLx3LN+dNJTpQgqQ9W7LjFd6ncMxc17OYPt8JQK5QNG9wIcojMScUTaFQFCihzrD24/fCbBGiipR4hmElY2MaZURV+UNbio3N+9jU3M4nbSkA6uMRTjh6BJMb6xjfkBi4wq0gPu3y+eHTa1m2ehvTx+aRIGnbKnjlF7D2KRfgbeYlMOc61/3PBG521CfroHl9L6GYAQ2TLTe3URqG4WB3ubBfdB+k0gHbWjrYuKudzbvbaU8FiMBRI+OcNa2RyY1JGutiB7dmoQLITZD0zXlT+VZ/CZKCNKx9wo1XfLTaRRade4Mbr6g/Clo3w/pfO9eT3+kWTPW4nkwoDONwxn7dIa0dnluJvaudj1o7CTJKTbSKSWOSPWsnCpIjogzknSCps9W5n1bdB3u3wegpcN4dLlZPx27Y+baLNNojFMc619PoKSYUhjFMGLZjGkFG2dHaGS6ya6el3a2dGFMX6xGKo0cNvHailByq7/r1LXu4cflbAydIatnkxiv+b6nLYjbxbLcY74gT3PhE83rwOnKEYkYoFOUJyWB+/Cxmiyxmiyw2plEgOr2gRyS2tLST8jNUVwnjGxLMGj+SyY3JQ4/aWmGk0gF3/vZ9lrwQJki65kzmTstJkKTqQiysXOxCfUgVzPxjt1VVuQHtnWtcD2LMsc71VEahMAyjMjisRUNVad7n9YTs+HivCwCYrKlmWlMdU5qSTBhd27dffwiTmyDp8tkTuOWCnARJge8WQ72yGHa84dZXnPbnMH62S4H58ZuhUExzrqcxU00oDMPo4bATDT/IsK2lo6dH8WlXGnABAM+cPIYpTUmOqK8ZcoPY+eAHGe753Yfc/T/v05CM8cBVszlnxlj3YXesnleXQNtHLn/F6de4ge3Ad+dGTw17FFPdLCjDMIxeHBai0dblszkUia27O3oCAE4YXcucKW7tRF3NYfGq/ZKbIOnCk12CpIZkzEWMXXkvvLEU0h1wxEyY8kUYMc71IEZPMaEwDCNvhmRNmskoO9u6elZiN3/q1k6MTEQ5cbyL6zRuVILIEFk7MRiCjPLAS5v46X+vJxmrZvEVp3L+SUfClpfhycWw/hmQajjyJBfxc8Q4GDMldD1NM6EwDOOgGDKi0eUHbM1ZO9HpuQCAR4+K8/lj3dqJ0cmht3ZiMGzZ3c53l7/Nqs0tfPn4sdx20QyatjwD99wJn7wHkYTLnz3hDDhylutRjJlmmcsMwzhkKl40Xt/SwsZd7exo7SKjSjxazeTGWiY31jFxTGUGACw2qsp/vLqVHz+7lmoR7r7wGL7W/iiy5DIXvC0xGo6bD8dfCEefbEJhGEbBqHjR+P2GZhrra5g9qYFJjUmOGlG5AQBLQTZB0i6unrCT79Y9R/y3z0PGd8mKTvkzlwe5aboJhWEYBafiF/e1dniHlpP6MENEeOS1rfzy6Rc5N3iJa0euZNS+DxGphmPOhNOvhennufwUhzm2iCuL2SKL2SLLsF7cd9gIhqoL8KcBZNLh5vY1kyaV8tnX1UlnV4qOLo/OLo+OVIpUKkWn51arVz2+iEeia2mo3gOpJJx6FXz+O9DQT75twzCMAlPxojFoVEEz+1XSPft9VOCfuSaT7ve6TJCmy0vRlfJIeT6elyLleXi+j+95eH6atN9FJtWJ+p1k0h6ZtIcGHqR9yHiQ8anK+EQ0TVTSRAmIkiZGmrpwPyJurcklkZfRkRNh7k3ODRU9iLzdhmEYBaDkoiEiXwXuAqqB+1X19gFv2PhCPxX4ASr83OtyuqyKEmQUL51xW5Ah5WfwfY+0nyLwPYK0R8ZPuUo+8NC0j2Z8NHC9AjIBkvGpygREJU2EgBhpovjUkyYmaWrwqcEnKsGBjZIzMzhdFScTiZOJJiCSRGJJJF4HbIYrH0WmnuPCfBiGYZSBko5piEg1sAH4CrAdeA1YqKrv9XO96vO3QVU1GYngBYrnp/F815r3PJ902ift+wR+inTaJ0h7qO8qe4IUkvaQwKNKPSIZt9XgUYNPXDzipIjjUX0QY+sB1XjVtaSrEwSRWjKRBBpNQixJVY3bIvE6IvE6ovE6qmrqIFYLsTqI1va/H63tVxDMX5vFbJHFbJHFbJHlcBrTOAP4QFU3AojIQ8BFQJ+iAbDzd/9KXLtI0EVc0hzMMG8ncbyqOF5VAj+SIKhOEEQayERr8WJJvFgtn8bqqI4nicTriSbqiCXqiSfricbrkJgTAqLJsHJ3+9WRGOYYMgxjOFJq0RgHbMs53g6cmXuBiCwCFvVcMOp0MrEkEq1FapJU1dSFrfgksUQ9sdp64rX1JJIjSSTricTrXQUfSZCoqrLK3TAMo4CUWjT6cgLt159U1fuA+8C5p2Z/66FSlMswDMPIg1KPqG4HJuQcjwd2lLgMhmEYxiFSatF4DThWRCaLSAxYADxZ4jIYhmEYh0hJ3VOqmhaRvwKew025fUBV3y1lGQzDMIxDp+LDiFRy+UqJTSfMYrbIYrbIYrbIUswpt7ZKzDAMw8gbEw3DMAwjb0w0DMMwjLwx0TAMwzDyptKj3O4TkfXlLkSF0CgizeUuRIVgtshitshitsgyvVgPrnTRWF+sGQBDDRFZbbZwmC2ymC2ymC2yiMjqYj3b3FOGYRhG3phoGIZhGHlT6aJxX7kLUEGYLbKYLbKYLbKYLbIUzRYVvSLcMAzDqCwqvadhGIZhVBAlEw0RURFZmnMcEZFdIvJ0eHyViPyij/t+JCLbRGRfr/O3hs+clnPu2+G5iptBUYT3rxGRZSLygYi8KiKTwvPzwr91dc61p4TnbizW+xUDEQlE5E0ReUtE3hCRueH52SLyThgpGRGZKiIbRWREeUt8aIjIJeH/z4zweFJ4/E851zSKiN/9HRGR50LbdG87ROTVcr1DMRjOdhHHSyIyP+fc10VkxQD3bBeR53ude0dE3ixk2UrZ02gHThSR7mR6XwE+yuO+p3BpYvtiDS68ejeXMkDq2DJT6Pe/GtijqtOAnwP/nPPZGuDynOMFwFsHXeLy06mqn1PVk4G/B24DUNXVwO+BbhFcDNysqm3lKeagWQi8xP7f5Y3ABTnHlwE9EaFV9dzQNp8DzgLagFtKUNZSMmztEkZq/UvgX0QkLiJJ4EfA9Qe4dZSIHA0gIicB6UKXrdTuqV8D54f7C4H/PNANqrpSVT/u5+PHcTnGEZEpwF5gVwHKWSwK+f4XAf8e7j8CfElEujMjbgXiIjI2PPfV8G8PZUYAe3KObwKuEZG/A6KqekBbViIiUoer3K5m/8qxE1ib02u+HHi4n8fcBTyrqr8RkeNFZFXO8yeJyNvh/nkisi5swd7d3cutRAptl/CZU0VkhYi8LiIvdvdgKhVVfQfXaPwe8A/Ag6r6oYg8Fb7DuyJyTa/blgNfD/f3q2NE5BoRuTPneIWInB16PZaKyJqwZ3LDQOUqtWg8BCwQkTgwCxhst7EN2CYiJ+IMtGyQzys2hXz/nnzrqprGCeaYnM8fwbXC5gJvAKlB/K1ykQhdDOuA+4Eet4SqtuJ6V7cB15WpfIXgYmCFqm4AWkTk1JzPur8v44GAPrJcisglwGxcTwxVXQvEwkYUhJVq+J1bAsxX1bOBpmK9UIEoqF1C7gP+WlVPw/VSf1mswheQfwSuAOYDPwnPfSN8h9OBvxWRhpzrl+M8LgDnAc/k8TdOAxpV9SRVPRF4cKCLSyoaqvo2MAlXwT9boMc+hGuJXAw8VqBnFoUCv/+B8q0/jBONvHo0FUq3e2oGrrf0YE5vCtwP6Q/ACWUpXWFYiPsOE/67MOezFTg3Zp8NIhEZB9wNXKGquY2Ch8m2Ni8P750BbFTVTeH5Sv9OFNQuYc9lLrA89PEvAY4qWukLhKq2495xac7/8bdF5C3gFVzK7Kk5t+wC2kVkAfA20JXHn/kAmC4id4nIubgGaL+UI4zIk8AdwDz2bxkfKk8BPwVWq2rb/nVKRVKo9+/Ot75dRCLASKCl+0NV3SkiPu7H9Te4H8yQRVVfEZFGXAv5ExG5APfO5wKPichzqtpR1kIeJCIyBjgHN9aluGyWStgCVlVPRF4HvgPMBL6Wc6/g3JO3q2rvcbxluMrxv9xj9H0ROaXoL1QgimSXKqA1HOsYamTCDRH5MvAFYI6qdorIS0C81/XLcON8f9rrfJr9OwpxAFXdLSKzcI2wG4A/ARb1V5hyiMYDwF5VXSMi8wb7sNBw3wM2DLpkpaFQ7/8k8A1ca+NS4H9VVXuJ5g+AI1Q1GAJiOiCh/7ka2B1OJvgZcImqviciTwA3h9tQ4lKcn/ovuk+IyAu41mM3PwNeCH/YuffeCHSp6uLeDw393gHwfbIt8XXAFBGZpKqb2X+iRKVRcLuEDcpNInKZqi4PxWWWqg61CSIjgZaw3puJc1H15lFc4+o3OM9GN5uBq8N3n4hzSyEiTTibLReRTcC9AxWg5KKhqttxA1R9cZWIXJxzPAenfFcAtSKyHbhfVW/t9cyHGCIU8P1/BSwVkQ9wPYwFvR+mqi8XsuxlIJEzXVBwvtxARL4PPJ7TkrwVeFNE/k1V3y9HQQ+RhcDtvc49ihvkB0BV3yVndlAOP8T1MnOnU+5R1S+G+8twPfDJ4XM6ReQ6YIW4SLCrqFyKZZcrgXtE5BYginN7DTXReAZYFLqn1tHHuKiq7iWcTdlLUF/AzdhcA7wDdNtoAvCrUEwUN/DeL7Yi3DCGCSJSp6r7wsphMfC+qv683OUyhha2Itwwhg/Xhi3wd3FujiVlLo8xBLGehmEYhpE31tMwDMMw8sZEwzAMw8gbEw3DMAwjb0w0DMMwjLwx0TAMwzDyxkTDMAzDyJv/B5yGFjLhSIWaAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('data').loc['HybridSVD(cs)', 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'scaling', color=colors)\n",
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('data').loc['PureSVD(cs)', 'rank'].loc[:, data_labels].reset_index(),\n",
    "                     'scaling', color=colors, alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>scaling</th>\n",
       "      <th>False</th>\n",
       "      <th>True</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</th>\n",
       "      <th>data</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">HybridSVD(cs)</th>\n",
       "      <th>AMZe</th>\n",
       "      <td>1750</td>\n",
       "      <td>750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMZvg</th>\n",
       "      <td>2000</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BX</th>\n",
       "      <td>750</td>\n",
       "      <td>500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML10M</th>\n",
       "      <td>60</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML1M</th>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     rank      \n",
       "scaling             False True \n",
       "model         data             \n",
       "HybridSVD(cs) AMZe   1750   750\n",
       "              AMZvg  2000   500\n",
       "              BX      750   500\n",
       "              ML10M    60    50\n",
       "              ML1M      5     5"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.set_index(['model','scaling','data']).sort_index().unstack('scaling').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x216c2fd1c18>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('scaling').loc['HybridSVD(cs)', 'rank'].reset_index().dropna(),\n",
    "                     'data', color=colors)\n",
    "parallel_coordinates(data.set_index(['model','scaling','data']).sort_index().unstack('scaling').loc['PureSVD(cs)', 'rank'].reset_index().dropna(),\n",
    "                     'data', color=colors, alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">rank</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>model</th>\n",
       "      <th>HybridSVD(cs)</th>\n",
       "      <th>PureSVD(cs)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scaling</th>\n",
       "      <th>data</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">False</th>\n",
       "      <th>AMZe</th>\n",
       "      <td>1750</td>\n",
       "      <td>2500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMZvg</th>\n",
       "      <td>2000</td>\n",
       "      <td>3000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BX</th>\n",
       "      <td>750</td>\n",
       "      <td>400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML10M</th>\n",
       "      <td>60</td>\n",
       "      <td>400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ML1M</th>\n",
       "      <td>5</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       rank            \n",
       "model         HybridSVD(cs) PureSVD(cs)\n",
       "scaling data                           \n",
       "False   AMZe           1750        2500\n",
       "        AMZvg          2000        3000\n",
       "        BX              750         400\n",
       "        ML10M            60         400\n",
       "        ML1M              5         100"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.set_index(['model','scaling','data']).sort_index().unstack('model').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}