RNN-3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from __future__ import print_function\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Input, LSTM, Dense, Dropout\n",
    "import numpy as np\n",
    "import csv\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import r2_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "nteams = 20\n",
    "nmatches = (nteams-1)*2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "batch_size = 64  # Batch size for training.\n",
    "epochs = 100  # Number of epochs to train for.\n",
    "latent_dim = 256  # Latent dimensionality of the encoding space.\n",
    "num_samples = 10000  # Number of samples to train on.\n",
    "# Path to the data txt file on disk.\n",
    "data_path = 'fra-eng/fra.txt'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data Pre-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def season_team_map(season_dir):\n",
    "    readme = season_dir + '/README.md'\n",
    "    f = open(readme, 'r')\n",
    "    \n",
    "    s = \"  \"\n",
    "    \n",
    "    # skip to teams\n",
    "    while(s[1] != '1'):\n",
    "        s = f.readline()\n",
    "        print(s)\n",
    "        if(len(s)<=1):\n",
    "            s = \"  \"\n",
    "\n",
    "    team_map = {}\n",
    "\n",
    "    for i in range(20):\n",
    "        team_map[s[4:].split('  ')[0]] = i\n",
    "        s = f.readline()\n",
    "\n",
    "    f.close()\n",
    "    \n",
    "    print(team_map)\n",
    "    return team_map"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "def season_team_map_div_2(season_dir):\n",
    "    readme = season_dir + '/README.md'\n",
    "    f = open(readme, 'r')\n",
    "    \n",
    "    s = \"  \"\n",
    "    \n",
    "    # skip to teams\n",
    "    while(s[1] != '1'):\n",
    "        s = f.readline()\n",
    "        if(len(s)<=1):\n",
    "            s = \"  \"\n",
    "            \n",
    "    s = f.readline()\n",
    "    if(len(s)<=1):\n",
    "        s = \"  \"    \n",
    "        \n",
    "     # skip to div 2 teams\n",
    "    while(s[:3] != ' 1.'):\n",
    "        s = f.readline()\n",
    "        if(len(s)<=1):\n",
    "            s = \"  \"\n",
    "\n",
    "    team_map = {}\n",
    "\n",
    "    for i in range(20):\n",
    "        team_map[s[4:].split('  ')[0]] = i\n",
    "        s = f.readline()\n",
    "\n",
    "    f.close()\n",
    "    \n",
    "    print(team_map)\n",
    "    return team_map"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "def season_game_mat(season_dir, team_map, div=1):\n",
    "    game_mat = np.zeros((nteams, nmatches), dtype=int)\n",
    "    team_index = np.zeros(nteams, dtype=int)\n",
    "\n",
    "    csvfile = ''\n",
    "    print(div)\n",
    "    \n",
    "    for entry in os.scandir(season_dir):\n",
    "        if(div==1):\n",
    "            if '/1-' in entry.path:\n",
    "                csvfile = entry.path\n",
    "        elif(div==2):\n",
    "            if '/2-' in entry.path:\n",
    "                csvfile = entry.path\n",
    "            \n",
    "    with open(csvfile) as gf:\n",
    "        reader = csv.reader(gf)\n",
    "        header = next(reader)\n",
    "        if(header[0] == 'Round'):\n",
    "            new_format = 1\n",
    "        else:\n",
    "            new_format = 0;\n",
    "        for row in reader:\n",
    "            if(new_format):\n",
    "                team1 = team_map[row[2].split(' (')[0]]\n",
    "                team2 = team_map[row[5].split(' (')[0]]\n",
    "\n",
    "                score = row[3].split('-')\n",
    "            else:\n",
    "                team1 = team_map[row[1]]\n",
    "                team2 = team_map[row[2]]\n",
    "\n",
    "                score = row[3].split('-')\n",
    "            \n",
    "\n",
    "            # 1 for win, 0 for draw, -1 for loss\n",
    "\n",
    "            if(int(score[0]) > int(score[1])):\n",
    "                game_mat[team1, team_index[team1]] = 1\n",
    "                game_mat[team2, team_index[team2]] = -1\n",
    "            elif(int(score[0]) < int(score[1])):\n",
    "                game_mat[team1, team_index[team1]] = -1\n",
    "                game_mat[team2, team_index[team2]] = 1\n",
    "\n",
    "            team_index[team1] += 1\n",
    "            team_index[team2] += 1\n",
    "    \n",
    "    return np.transpose(game_mat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./data/fr-france-master\n",
      "2017-18\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Paris Saint-Germain           38  17  1  1  70:15   12  5  2  38:14   108:29  +79   93\n",
      "\n",
      "{'Paris Saint-Germain': 0, 'AS Monaco': 1, 'Olympique Lyonnais': 2, 'Olympique de Marseille': 3, 'Stade Rennais FC': 4, 'Girondins de Bordeaux': 5, 'AS Saint-Étienne': 6, 'OGC Nice': 7, 'FC Nantes': 8, 'Montpellier HSC': 9, 'Dijon FCO': 10, 'EA Guingamp': 11, 'Amiens SC': 12, 'Angers SCO': 13, 'RC Strasbourg': 14, 'Stade Malherbe Caen': 15, 'Lille OSC': 16, 'Toulouse FC': 17, 'ES Troyes AC': 18, 'FC Metz': 19}\n",
      "1\n",
      "{'Stade de Reims': 0, 'Nîmes Olympique': 1, 'AC Ajaccio': 2, 'Le Havre AC': 3, 'Stade Brestois 29': 4, 'Clermont Foot 63': 5, 'FC Lorient': 6, 'Paris FC': 7, 'LB Châteauroux': 8, 'FC Sochaux-Montbéliard': 9, 'AJ Auxerre': 10, 'US Orléans': 11, 'Valenciennes FC': 12, 'RC Lens': 13, 'Chamois Niortais FC': 14, 'GFC Ajaccio': 15, 'AS Nancy Lorraine': 16, 'Bourg-en-Bresse 01': 17, 'US Quevilly-Rouen Métropole': 18, 'Tours FC': 19}\n",
      "2\n",
      "1995-96\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Auxerre                       38  22   6  10  66:30   72\n",
      "\n",
      "{'Auxerre': 0, 'Paris SG': 1, 'Monaco': 2, 'Metz': 3, 'Lens': 4, 'Montpellier': 5, 'Nantes': 6, 'Rennes': 7, 'Strasbourg': 8, 'Guingamp': 9, 'Lyon': 10, 'Nice': 11, 'Le Havre': 12, 'Cannes': 13, 'Bastia': 14, 'Bordeaux': 15, 'Lille': 16, 'Gueugnon': 17, 'St Etienne': 18, 'Martigues': 19}\n",
      "1\n",
      ".DS_Store\n",
      "2016-17\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. AS Monaco                     38  17  1  1  63:13   13  4  2  44:18   107:31  +76   95\n",
      "\n",
      "{'AS Monaco': 0, 'Paris Saint-Germain': 1, 'OGC Nice': 2, 'Olympique Lyonnais': 3, 'Olympique de Marseille': 4, 'Girondins de Bordeaux': 5, 'FC Nantes': 6, 'AS Saint-Étienne': 7, 'Stade Rennais FC': 8, 'EA Guingamp': 9, 'Lille OSC': 10, 'Angers SCO': 11, 'Toulouse FC': 12, 'FC Metz': 13, 'Montpellier HSC': 14, 'Dijon FCO': 15, 'Stade Malherbe Caen': 16, 'FC Lorient': 17, 'AS Nancy Lorraine': 18, 'SC Bastia': 19}\n",
      "1\n",
      "{'RC Strasbourg': 0, 'Amiens SC': 1, 'ES Troyes AC': 2, 'RC Lens': 3, 'Stade Brestois 29': 4, 'Nîmes Olympique': 5, 'Stade de Reims': 6, 'Le Havre AC': 7, 'GFC Ajaccio': 8, 'Chamois Niortais FC': 9, 'AC Ajaccio': 10, 'Clermont Foot 63': 11, 'FC Sochaux-Montbéliard': 12, 'Valenciennes FC': 13, 'Bourg-en-Bresse 01': 14, 'Tours FC': 15, 'AJ Auxerre': 16, 'US Orléans': 17, 'Red Star FC': 18, 'Stade Lavallois': 19}\n",
      "2\n",
      "SOURCES.md\n",
      "2005-06\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lyon                          38  25   9   4  73:31   84\n",
      "\n",
      "{'Lyon': 0, 'Bordeaux': 1, 'Lille': 2, 'Lens': 3, 'Marseille': 4, 'Auxerre': 5, 'Rennes': 6, 'Nice': 7, 'Paris SG': 8, 'Monaco': 9, 'Le Mans': 10, 'Nancy': 11, 'St Etienne': 12, 'Nantes': 13, 'Sochaux': 14, 'Toulouse': 15, 'Troyes': 16, 'Ajaccio': 17, 'Strasbourg': 18, 'Metz': 19}\n",
      "1\n",
      "{'Valenciennes': 0, 'Sedan': 1, 'Lorient': 2, 'Caen': 3, 'Dijon': 4, 'Bastia': 5, 'Le Havre': 6, 'Creteil': 7, 'Guingamp': 8, 'Grenoble': 9, 'Gueugnon': 10, 'Montpellier': 11, 'Istres': 12, 'Reims': 13, 'Chateauroux': 14, 'Amiens': 15, 'Brest': 16, 'Clermont': 17, 'Laval': 18, 'Sete': 19}\n",
      "2\n",
      "1993-94\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Paris SG                      38  24  11   3  54:22   83\n",
      "\n",
      "{'Paris SG': 0, 'Marseille': 1, 'Bordeaux': 2, 'Auxerre': 3, 'Nantes': 4, 'Cannes': 5, 'Lyon': 6, 'Montpellier': 7, 'Monaco': 8, 'Lens': 9, 'St Etienne': 10, 'Metz': 11, 'Strasbourg': 12, 'Sochaux': 13, 'Caen': 14, 'Lille': 15, 'Le Havre': 16, 'Martigues': 17, 'Toulouse': 18, 'Angers': 19}\n",
      "1\n",
      "2002-03\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lyon                          38  19  11   8  63:41   68\n",
      "\n",
      "{'Lyon': 0, 'Monaco': 1, 'Marseille': 2, 'Bordeaux': 3, 'Sochaux': 4, 'Auxerre': 5, 'Guingamp': 6, 'Lens': 7, 'Nantes': 8, 'Nice': 9, 'Paris SG': 10, 'Bastia': 11, 'Strasbourg': 12, 'Lille': 13, 'Rennes': 14, 'Montpellier': 15, 'Ajaccio': 16, 'Le Havre': 17, 'Sedan': 18, 'Troyes': 19}\n",
      "1\n",
      "{'Toulouse': 0, 'Le Mans': 1, 'Metz': 2, 'Lorient': 3, 'Chateauroux': 4, 'Caen': 5, 'Laval': 6, 'St Etienne': 7, 'Niort': 8, 'Amiens': 9, 'Gueugnon': 10, 'Grenoble': 11, 'Clermont': 12, 'Valence': 13, 'Nancy': 14, 'Istres': 15, 'Creteil': 16, 'Beauvais': 17, 'Wasquehal': 18, 'Reims': 19}\n",
      "2\n",
      "2011-12\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Montpellier                   38  25   7   6  68:34   82\n",
      "\n",
      "{'Montpellier': 0, 'Paris SG': 1, 'Lille': 2, 'Lyon': 3, 'Bordeaux': 4, 'Rennes': 5, 'St Etienne': 6, 'Toulouse': 7, 'Evian Thonon Gaillard': 8, 'Marseille': 9, 'Nancy': 10, 'Valenciennes': 11, 'Nice': 12, 'Sochaux': 13, 'Brest': 14, 'Ajaccio': 15, 'Lorient': 16, 'Caen': 17, 'Dijon': 18, 'Auxerre': 19}\n",
      "1\n",
      "{'Bastia': 0, 'Reims': 1, 'Troyes': 2, 'Sedan': 3, 'Clermont': 4, 'Tours': 5, 'Guingamp': 6, 'Monaco': 7, 'Nantes': 8, 'Istres': 9, 'Angers': 10, 'Lens': 11, 'Arles': 12, 'Chateauroux': 13, 'Le Havre': 14, 'Laval': 15, 'Le Mans': 16, 'Metz': 17, 'Boulogne': 18, 'Amiens': 19}\n",
      "2\n",
      "2009-10\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Marseille                     38  23   9   6  69:36   78\n",
      "\n",
      "{'Marseille': 0, 'Lyon': 1, 'Auxerre': 2, 'Lille': 3, 'Montpellier': 4, 'Bordeaux': 5, 'Lorient': 6, 'Monaco': 7, 'Rennes': 8, 'Valenciennes': 9, 'Lens': 10, 'Nancy': 11, 'Paris SG': 12, 'Toulouse': 13, 'Nice': 14, 'Sochaux': 15, 'St Etienne': 16, 'Le Mans': 17, 'Boulogne': 18, 'Grenoble': 19}\n",
      "1\n",
      "{'Caen': 0, 'Brest': 1, 'Arles': 2, 'Metz': 3, 'Angers': 4, 'Clermont': 5, 'Le Havre': 6, 'Laval': 7, 'Dijon': 8, 'Nimes': 9, 'Tours': 10, 'Sedan': 11, 'Ajaccio': 12, 'Vannes': 13, 'Nantes': 14, 'Chateauroux': 15, 'Istres': 16, 'Guingamp': 17, 'Strasbourg': 18, 'Bastia': 19}\n",
      "2\n",
      "2003-04\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lyon                          38  24   7   7  64:26   79\n",
      "\n",
      "{'Lyon': 0, 'Paris SG': 1, 'Monaco': 2, 'Auxerre': 3, 'Sochaux': 4, 'Nantes': 5, 'Marseille': 6, 'Lens': 7, 'Rennes': 8, 'Lille': 9, 'Nice': 10, 'Bordeaux': 11, 'Strasbourg': 12, 'Metz': 13, 'Ajaccio': 14, 'Toulouse': 15, 'Bastia': 16, 'Guingamp': 17, 'Le Mans': 18, 'Montpellier': 19}\n",
      "1\n",
      "{'St Etienne': 0, 'Caen': 1, 'Istres': 2, 'Lorient': 3, 'Sedan': 4, 'Nancy': 5, 'Le Havre': 6, 'Niort': 7, 'Amiens': 8, 'Troyes': 9, 'Chateauroux': 10, 'Creteil': 11, 'Angers': 12, 'Clermont': 13, 'Grenoble': 14, 'Gueugnon': 15, 'Laval': 16, 'Valence': 17, 'Besancon': 18, 'Rouen': 19}\n",
      "2\n",
      "2014-15\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Paris Saint-Germain           38  15  4  0  52:14    9  7  3  31:22    83:36  +47   83\n",
      "\n",
      "{'Paris Saint-Germain': 0, 'Olympique Lyonnais': 1, 'AS Monaco': 2, 'Olympique de Marseille': 3, 'AS Saint-Étienne': 4, 'Girondins de Bordeaux': 5, 'Montpellier HSC': 6, 'Lille OSC': 7, 'Stade Rennais FC': 8, 'EA Guingamp': 9, 'OGC Nice': 10, 'SC Bastia': 11, 'Stade Malherbe Caen': 12, 'FC Nantes': 13, 'Stade de Reims': 14, 'FC Lorient': 15, 'Toulouse FC': 16, 'Évian TG FC': 17, 'FC Metz': 18, 'RC Lens': 19}\n",
      "1\n",
      "{'ES Troyes AC': 0, 'GFC Ajaccio': 1, 'Angers SCO': 2, 'Dijon FCO': 3, 'AS Nancy Lorraine': 4, 'Stade Brestois 29': 5, 'Le Havre AC': 6, 'Stade Lavallois': 7, 'AJ Auxerre': 8, 'FC Sochaux-Montbéliard': 9, 'Chamois Niortais FC': 10, 'Clermont Foot 63': 11, 'Nîmes Olympique': 12, 'US Créteil': 13, 'Tours FC': 14, 'Valenciennes FC': 15, 'AC Ajaccio': 16, 'US Orléans': 17, 'LB Châteauroux': 18, 'AC Arles-Avignon': 19}\n",
      "2\n",
      "README.md\n",
      "2004-05\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lyon                          38  22  13   3  56:22   79\n",
      "\n",
      "{'Lyon': 0, 'Lille': 1, 'Monaco': 2, 'Rennes': 3, 'Marseille': 4, 'St Etienne': 5, 'Lens': 6, 'Auxerre': 7, 'Paris SG': 8, 'Sochaux': 9, 'Strasbourg': 10, 'Nice': 11, 'Toulouse': 12, 'Ajaccio': 13, 'Bordeaux': 14, 'Metz': 15, 'Nantes': 16, 'Caen': 17, 'Bastia': 18, 'Istres': 19}\n",
      "1\n",
      "{'Nancy': 0, 'Le Mans': 1, 'Troyes': 2, 'Dijon': 3, 'Chateauroux': 4, 'Sedan': 5, 'Guingamp': 6, 'Montpellier': 7, 'Brest': 8, 'Lorient': 9, 'Grenoble': 10, 'Gueugnon': 11, 'Amiens': 12, 'Laval': 13, 'Creteil': 14, 'Reims': 15, 'Le Havre': 16, 'Clermont': 17, 'Angers': 18, 'Niort': 19}\n",
      "2\n",
      "2008-09\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Bordeaux                      38  24   8   6  64:34   80\n",
      "\n",
      "{'Bordeaux': 0, 'Marseille': 1, 'Lyon': 2, 'Toulouse': 3, 'Lille': 4, 'Paris SG': 5, 'Rennes': 6, 'Auxerre': 7, 'Nice': 8, 'Lorient': 9, 'Monaco': 10, 'Valenciennes': 11, 'Grenoble': 12, 'Sochaux': 13, 'Nancy': 14, 'Le Mans': 15, 'St Etienne': 16, 'Caen': 17, 'Nantes': 18, 'Le Havre': 19}\n",
      "1\n",
      "{'Lens': 0, 'Montpellier': 1, 'Boulogne': 2, 'Strasbourg': 3, 'Metz': 4, 'Tours': 5, 'Angers': 6, 'Dijon': 7, 'Sedan': 8, 'Vannes': 9, 'Bastia': 10, 'Clermont': 11, 'Guingamp': 12, 'Chateauroux': 13, 'Ajaccio': 14, 'Nimes': 15, 'Amiens': 16, 'Brest': 17, 'Reims': 18, 'Troyes': 19}\n",
      "2\n",
      "2013-14\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Paris SG                      38  27   8   3  84:23   89\n",
      "\n",
      "{'Paris SG': 0, 'Monaco': 1, 'Lille': 2, 'St Etienne': 3, 'Lyon': 4, 'Marseille': 5, 'Bordeaux': 6, 'Nantes': 7, 'Lorient': 8, 'Toulouse': 9, 'Reims': 10, 'Rennes': 11, 'Bastia': 12, 'Evian Thonon Gaillard': 13, 'Montpellier': 14, 'Guingamp': 15, 'Nice': 16, 'Sochaux': 17, 'Valenciennes': 18, 'Ajaccio': 19}\n",
      "1\n",
      "{'Metz': 0, 'Lens': 1, 'Caen': 2, 'Nancy': 3, 'Niort': 4, 'Dijon': 5, 'Brest': 6, 'Tours': 7, 'Angers': 8, 'Troyes': 9, 'Creteil': 10, 'Le Havre': 11, 'Arles': 12, 'Clermont': 13, 'Nimes': 14, 'Auxerre': 15, 'Laval': 16, 'Chateauroux': 17, 'Istres': 18, 'CA Bastia': 19}\n",
      "2\n",
      "1994-95\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Nantes                        38  21  16   1  71:34   79\n",
      "\n",
      "{'Nantes': 0, 'Lyon': 1, 'Paris SG': 2, 'Auxerre': 3, 'Lens': 4, 'Monaco': 5, 'Bordeaux': 6, 'Metz': 7, 'Cannes': 8, 'Strasbourg': 9, 'Martigues': 10, 'Le Havre': 11, 'Rennes': 12, 'Lille': 13, 'Bastia': 14, 'Nice': 15, 'Montpellier': 16, 'St Etienne': 17, 'Caen': 18, 'Sochaux': 19}\n",
      "1\n",
      "2012-13\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Paris SG                      38  25   8   5  69:23   83\n",
      "\n",
      "{'Paris SG': 0, 'Marseille': 1, 'Lyon': 2, 'Nice': 3, 'St Etienne': 4, 'Lille': 5, 'Bordeaux': 6, 'Lorient': 7, 'Montpellier': 8, 'Toulouse': 9, 'Valenciennes': 10, 'Bastia': 11, 'Rennes': 12, 'Reims': 13, 'Ajaccio': 14, 'Sochaux': 15, 'Evian Thonon Gaillard': 16, 'Nancy': 17, 'Troyes': 18, 'Brest': 19}\n",
      "1\n",
      "{'Monaco': 0, 'Guingamp': 1, 'Nantes': 2, 'Caen': 3, 'Angers': 4, 'Le Havre': 5, 'Dijon': 6, 'Nimes': 7, 'Auxerre': 8, 'Tours': 9, 'Arles': 10, 'Lens': 11, 'Istres': 12, 'Clermont': 13, 'Niort': 14, 'Chateauroux': 15, 'Laval': 16, 'Le Mans': 17, 'Sedan': 18, 'Ajaccio GFCO': 19}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\n",
      "2015-16\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Paris Saint-Germain           38  15  3  1  59:12   15  3  1  43:7    102:19  +83   96\n",
      "\n",
      "{'Paris Saint-Germain': 0, 'Olympique Lyonnais': 1, 'AS Monaco': 2, 'OGC Nice': 3, 'Lille OSC': 4, 'AS Saint-Étienne': 5, 'Stade Malherbe Caen': 6, 'Stade Rennais FC': 7, 'Angers SCO': 8, 'SC Bastia': 9, 'Girondins de Bordeaux': 10, 'Montpellier HSC': 11, 'Olympique de Marseille': 12, 'FC Nantes': 13, 'FC Lorient': 14, 'EA Guingamp': 15, 'Toulouse FC': 16, 'Stade de Reims': 17, 'GFC Ajaccio': 18, 'ES Troyes AC': 19}\n",
      "1\n",
      "{'AS Nancy Lorraine': 0, 'Dijon FCO': 1, 'FC Metz': 2, 'Le Havre AC': 3, 'Red Star FC': 4, 'RC Lens': 5, 'Clermont Foot 63': 6, 'AJ Auxerre': 7, 'Nîmes Olympique': 8, 'Tours FC': 9, 'Stade Brestois 29': 10, 'Bourg-en-Bresse 01': 11, 'Valenciennes FC': 12, 'Stade Lavallois': 13, 'FC Sochaux-Montbéliard': 14, 'Chamois Niortais FC': 15, 'AC Ajaccio': 16, 'Évian TG FC': 17, 'US Créteil': 18, 'Paris FC': 19}\n",
      "2\n",
      "2006-07\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lyon                          38  24   9   5  64:27   81\n",
      "\n",
      "{'Lyon': 0, 'Marseille': 1, 'Rennes': 2, 'Lens': 3, 'Bordeaux': 4, 'Sochaux': 5, 'Toulouse': 6, 'Auxerre': 7, 'Monaco': 8, 'Lille': 9, 'St Etienne': 10, 'Le Mans': 11, 'Nancy': 12, 'Lorient': 13, 'Paris SG': 14, 'Nice': 15, 'Valenciennes': 16, 'Troyes': 17, 'Sedan': 18, 'Nantes': 19}\n",
      "1\n",
      "{'Metz': 0, 'Caen': 1, 'Strasbourg': 2, 'Amiens': 3, 'Grenoble': 4, 'Le Havre': 5, 'Chateauroux': 6, 'Dijon': 7, 'Bastia': 8, 'Gueugnon': 9, 'Reims': 10, 'Ajaccio': 11, 'Guingamp': 12, 'Brest': 13, 'Montpellier': 14, 'Niort': 15, 'Libourne': 16, 'Creteil': 17, 'Istres': 18, 'Tours': 19}\n",
      "2\n",
      "2010-11\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lille                         38  21  13   4  68:36   76\n",
      "\n",
      "{'Lille': 0, 'Marseille': 1, 'Lyon': 2, 'Paris SG': 3, 'Sochaux': 4, 'Rennes': 5, 'Bordeaux': 6, 'Toulouse': 7, 'Auxerre': 8, 'St Etienne': 9, 'Lorient': 10, 'Valenciennes': 11, 'Nancy': 12, 'Montpellier': 13, 'Caen': 14, 'Brest': 15, 'Nice': 16, 'Monaco': 17, 'Lens': 18, 'Arles': 19}\n",
      "1\n",
      "{'Evian Thonon Gaillard': 0, 'Ajaccio': 1, 'Dijon': 2, 'Le Mans': 3, 'Sedan': 4, 'Angers': 5, 'Clermont': 6, 'Boulogne': 7, 'Le Havre': 8, 'Reims': 9, 'Istres': 10, 'Tours': 11, 'Nantes': 12, 'Chateauroux': 13, 'Laval': 14, 'Troyes': 15, 'Metz': 16, 'Vannes': 17, 'Nimes': 18, 'Grenoble': 19}\n",
      "2\n",
      "2007-08\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Lyon                          38  24   7   7  74:37   79\n",
      "\n",
      "{'Lyon': 0, 'Bordeaux': 1, 'Marseille': 2, 'Nancy': 3, 'St Etienne': 4, 'Rennes': 5, 'Lille': 6, 'Nice': 7, 'Le Mans': 8, 'Lorient': 9, 'Caen': 10, 'Monaco': 11, 'Valenciennes': 12, 'Sochaux': 13, 'Auxerre': 14, 'Paris SG': 15, 'Toulouse': 16, 'Lens': 17, 'Strasbourg': 18, 'Metz': 19}\n",
      "1\n",
      "{'Le Havre': 0, 'Nantes': 1, 'Grenoble': 2, 'Sedan': 3, 'Clermont': 4, 'Troyes': 5, 'Brest': 6, 'Montpellier': 7, 'Ajaccio': 8, 'Angers': 9, 'Bastia': 10, 'Guingamp': 11, 'Reims': 12, 'Amiens': 13, 'Chateauroux': 14, 'Boulogne': 15, 'Dijon': 16, 'Niort': 17, 'Libourne': 18, 'Gueugnon': 19}\n",
      "2\n",
      "1996-97\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Monaco                        38  23  10   5  69:30   79\n",
      "\n",
      "{'Monaco': 0, 'Paris SG': 1, 'Nantes': 2, 'Bordeaux': 3, 'Metz': 4, 'Auxerre': 5, 'Bastia': 6, 'Lyon': 7, 'Strasbourg': 8, 'Montpellier': 9, 'Marseille': 10, 'Guingamp': 11, 'Lens': 12, 'Le Havre': 13, 'Cannes': 14, 'Rennes': 15, 'Caen': 16, 'Nancy': 17, 'Lille': 18, 'Nice': 19}\n",
      "1\n",
      "./data/eng-england-master\n",
      "1995-96\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  15  4  0  36:9    10  3  6  37:26    73:35  +38   82\n",
      "\n",
      "{'Man United': 0, 'Newcastle': 1, 'Liverpool': 2, 'Aston Villa': 3, 'Arsenal': 4, 'Everton': 5, 'Blackburn': 6, 'Tottenham': 7, \"Nott'm Forest\": 8, 'West Ham': 9, 'Chelsea': 10, 'Middlesbrough': 11, 'Leeds': 12, 'Wimbledon': 13, 'Sheffield Weds': 14, 'Coventry': 15, 'Southampton': 16, 'Man City': 17, 'QPR': 18, 'Bolton': 19}\n",
      "1\n",
      "2000-01\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  15  2  2  49:12    9  6  4  30:19    79:31  +48   80\n",
      "\n",
      "{'Man United': 0, 'Arsenal': 1, 'Liverpool': 2, 'Leeds': 3, 'Ipswich': 4, 'Chelsea': 5, 'Sunderland': 6, 'Aston Villa': 7, 'Charlton': 8, 'Southampton': 9, 'Newcastle': 10, 'Tottenham': 11, 'Leicester': 12, 'Middlesbrough': 13, 'West Ham': 14, 'Everton': 15, 'Derby': 16, 'Man City': 17, 'Coventry': 18, 'Bradford': 19}\n",
      "1\n",
      ".DS_Store\n",
      "teams.csv\n",
      "debug.txt\n",
      "SOURCES.md\n",
      "2005-06\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Chelsea                       38  18  1  0  47:9    11  3  5  25:13    72:22  +50   91\n",
      "\n",
      "{'Chelsea': 0, 'Man United': 1, 'Liverpool': 2, 'Arsenal': 3, 'Tottenham': 4, 'Blackburn': 5, 'Newcastle': 6, 'Bolton': 7, 'West Ham': 8, 'Wigan': 9, 'Everton': 10, 'Fulham': 11, 'Charlton': 12, 'Middlesbrough': 13, 'Man City': 14, 'Aston Villa': 15, 'Portsmouth': 16, 'Birmingham': 17, 'West Brom': 18, 'Sunderland': 19}\n",
      "1\n",
      "2002-03\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  16  2  1  42:12    9  6  4  32:22    74:34  +40   83\n",
      "\n",
      "{'Man United': 0, 'Arsenal': 1, 'Newcastle': 2, 'Chelsea': 3, 'Liverpool': 4, 'Blackburn': 5, 'Everton': 6, 'Southampton': 7, 'Man City': 8, 'Tottenham': 9, 'Middlesboro': 10, 'Charlton': 11, 'Birmingham': 12, 'Fulham': 13, 'Leeds': 14, 'Aston Villa': 15, 'Bolton': 16, 'West Ham': 17, 'West Brom': 18, 'Sunderland': 19}\n",
      "1\n",
      "2011-12\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man City                      38  18  1  0  55:12   10  4  5  38:17    93:29  +64   89\n",
      "\n",
      "{'Man City': 0, 'Man United': 1, 'Arsenal': 2, 'Tottenham': 3, 'Newcastle': 4, 'Chelsea': 5, 'Everton': 6, 'Liverpool': 7, 'Fulham': 8, 'West Brom': 9, 'Swansea': 10, 'Norwich': 11, 'Sunderland': 12, 'Stoke': 13, 'Wigan': 14, 'Aston Villa': 15, 'QPR': 16, 'Bolton': 17, 'Blackburn': 18, 'Wolves': 19}\n",
      "1\n",
      "2009-10\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Chelsea                       38  17  1  1  68:14   10  4  5  35:18   103:32  +71   86\n",
      "\n",
      "{'Chelsea': 0, 'Man United': 1, 'Arsenal': 2, 'Tottenham': 3, 'Man City': 4, 'Aston Villa': 5, 'Liverpool': 6, 'Everton': 7, 'Birmingham': 8, 'Blackburn': 9, 'Stoke': 10, 'Fulham': 11, 'Sunderland': 12, 'Bolton': 13, 'Wolves': 14, 'Wigan': 15, 'West Ham': 16, 'Burnley': 17, 'Hull': 18, 'Portsmouth': 19}\n",
      "1\n",
      "2003-04\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Arsenal                       38  15  4  0  40:14   11  8  0  33:12    73:26  +47   90\n",
      "\n",
      "{'Arsenal': 0, 'Chelsea': 1, 'Man United': 2, 'Liverpool': 3, 'Newcastle': 4, 'Aston Villa': 5, 'Charlton': 6, 'Bolton': 7, 'Fulham': 8, 'Birmingham': 9, 'Middlesbrough': 10, 'Southampton': 11, 'Portsmouth': 12, 'Tottenham': 13, 'Blackburn': 14, 'Man City': 15, 'Everton': 16, 'Leicester': 17, 'Leeds': 18, 'Wolves': 19}\n",
      "1\n",
      "README.md\n",
      "2004-05\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Chelsea                       38  14  5  0  35:6    15  3  1  37:9     72:15  +57   95\n",
      "\n",
      "{'Chelsea': 0, 'Arsenal': 1, 'Man United': 2, 'Everton': 3, 'Liverpool': 4, 'Bolton': 5, 'Middlesbrough': 6, 'Man City': 7, 'Tottenham': 8, 'Aston Villa': 9, 'Charlton': 10, 'Birmingham': 11, 'Fulham': 12, 'Newcastle': 13, 'Blackburn': 14, 'Portsmouth': 15, 'West Brom': 16, 'Crystal Palace': 17, 'Norwich': 18, 'Southampton': 19}\n",
      "1\n",
      "1999-00\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  15  4  0  59:16   13  3  3  38:29    97:45  +52   91\n",
      "\n",
      "{'Man United': 0, 'Arsenal': 1, 'Leeds': 2, 'Liverpool': 3, 'Chelsea': 4, 'Aston Villa': 5, 'Sunderland': 6, 'Leicester': 7, 'West Ham': 8, 'Tottenham': 9, 'Newcastle': 10, 'Middlesbrough': 11, 'Everton': 12, 'Coventry': 13, 'Southampton': 14, 'Derby': 15, 'Bradford': 16, 'Wimbledon': 17, 'Sheffield Weds': 18, 'Watford': 19}\n",
      "1\n",
      "2008-09\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  16  2  1  43:13   12  4  3  25:11    68:24  +44   90\n",
      "\n",
      "{'Man United': 0, 'Liverpool': 1, 'Chelsea': 2, 'Arsenal': 3, 'Everton': 4, 'Aston Villa': 5, 'Fulham': 6, 'Tottenham': 7, 'West Ham': 8, 'Man City': 9, 'Wigan': 10, 'Stoke': 11, 'Bolton': 12, 'Portsmouth': 13, 'Blackburn': 14, 'Sunderland': 15, 'Hull': 16, 'Newcastle': 17, 'Middlesbrough': 18, 'West Brom': 19}\n",
      "1\n",
      "2013-14\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man City                      38  17  1  1  63:13   10  4  5  39:24   102:37  +65   86\n",
      "\n",
      "{'Man City': 0, 'Liverpool': 1, 'Chelsea': 2, 'Arsenal': 3, 'Everton': 4, 'Tottenham': 5, 'Man United': 6, 'Southampton': 7, 'Stoke': 8, 'Newcastle': 9, 'Crystal Palace': 10, 'Swansea': 11, 'West Ham': 12, 'Sunderland': 13, 'Aston Villa': 14, 'Hull': 15, 'West Brom': 16, 'Norwich': 17, 'Fulham': 18, 'Cardiff': 19}\n",
      "1\n",
      "2001-02\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Arsenal                       38  12  4  3  42:25   14  5  0  37:11    79:36  +43   87\n",
      "\n",
      "{'Arsenal': 0, 'Liverpool': 1, 'Man United': 2, 'Newcastle': 3, 'Leeds': 4, 'Chelsea': 5, 'West Ham': 6, 'Aston Villa': 7, 'Tottenham': 8, 'Blackburn': 9, 'Southampton': 10, 'Middlesbrough': 11, 'Fulham': 12, 'Charlton': 13, 'Everton': 14, 'Bolton': 15, 'Sunderland': 16, 'Ipswich': 17, 'Derby': 18, 'Leicester': 19}\n",
      "1\n",
      "2012-13\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  16  0  3  45:19   12  5  2  41:24    86:43  +43   89\n",
      "\n",
      "{'Man United': 0, 'Man City': 1, 'Chelsea': 2, 'Arsenal': 3, 'Tottenham': 4, 'Everton': 5, 'Liverpool': 6, 'West Brom': 7, 'Swansea': 8, 'West Ham': 9, 'Norwich': 10, 'Fulham': 11, 'Stoke': 12, 'Southampton': 13, 'Aston Villa': 14, 'Newcastle': 15, 'Sunderland': 16, 'Wigan': 17, 'Reading': 18, 'QPR': 19}\n",
      "1\n",
      "1998-99\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  14  4  1  45:18    8  9  2  35:19    80:37  +43   79\n",
      "\n",
      "{'Man United': 0, 'Arsenal': 1, 'Chelsea': 2, 'Leeds': 3, 'West Ham': 4, 'Aston Villa': 5, 'Liverpool': 6, 'Derby': 7, 'Middlesbrough': 8, 'Leicester': 9, 'Tottenham': 10, 'Sheffield Weds': 11, 'Newcastle': 12, 'Everton': 13, 'Coventry': 14, 'Wimbledon': 15, 'Southampton': 16, 'Charlton': 17, 'Blackburn': 18, \"Nott'm Forest\": 19}\n",
      "1\n",
      "2006-07\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  15  2  2  46:12   13  3  3  37:15    83:27  +56   89\n",
      "\n",
      "{'Man United': 0, 'Chelsea': 1, 'Liverpool': 2, 'Arsenal': 3, 'Tottenham': 4, 'Everton': 5, 'Bolton': 6, 'Reading': 7, 'Portsmouth': 8, 'Blackburn': 9, 'Aston Villa': 10, 'Middlesbrough': 11, 'Newcastle': 12, 'Man City': 13, 'West Ham': 14, 'Fulham': 15, 'Wigan': 16, 'Sheffield United': 17, 'Charlton': 18, 'Watford': 19}\n",
      "1\n",
      "1997-98\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Arsenal                       38  15  2  2  43:10    8  7  4  25:23    68:33  +35   78\n",
      "\n",
      "{'Arsenal': 0, 'Man United': 1, 'Liverpool': 2, 'Chelsea': 3, 'Leeds': 4, 'Blackburn': 5, 'Aston Villa': 6, 'West Ham': 7, 'Derby': 8, 'Leicester': 9, 'Coventry': 10, 'Southampton': 11, 'Newcastle': 12, 'Tottenham': 13, 'Wimbledon': 14, 'Sheffield Weds': 15, 'Everton': 16, 'Bolton': 17, 'Barnsley': 18, 'Crystal Palace': 19}\n",
      "1\n",
      "2010-11\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  18  1  0  49:12    5 10  4  29:25    78:37  +41   80\n",
      "\n",
      "{'Man United': 0, 'Chelsea': 1, 'Man City': 2, 'Arsenal': 3, 'Tottenham': 4, 'Liverpool': 5, 'Everton': 6, 'Fulham': 7, 'Aston Villa': 8, 'Sunderland': 9, 'West Brom': 10, 'Newcastle': 11, 'Stoke': 12, 'Bolton': 13, 'Blackburn': 14, 'Wigan': 15, 'Wolves': 16, 'Birmingham': 17, 'Blackpool': 18, 'West Ham': 19}\n",
      "1\n",
      "2007-08\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  17  1  1  47:7    10  5  4  33:15    80:22  +58   87\n",
      "\n",
      "{'Man United': 0, 'Chelsea': 1, 'Arsenal': 2, 'Liverpool': 3, 'Everton': 4, 'Aston Villa': 5, 'Blackburn': 6, 'Portsmouth': 7, 'Man City': 8, 'West Ham': 9, 'Tottenham': 10, 'Newcastle': 11, 'Middlesbrough': 12, 'Wigan': 13, 'Sunderland': 14, 'Bolton': 15, 'Fulham': 16, 'Reading': 17, 'Birmingham': 18, 'Derby': 19}\n",
      "1\n",
      "1996-97\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Man United                    38  12  5  2  38:17    9  7  3  38:27    76:44  +32   75\n",
      "\n",
      "{'Man United': 0, 'Newcastle': 1, 'Arsenal': 2, 'Liverpool': 3, 'Aston Villa': 4, 'Chelsea': 5, 'Sheffield Weds': 6, 'Wimbledon': 7, 'Leicester': 8, 'Tottenham': 9, 'Leeds': 10, 'Derby': 11, 'Blackburn': 12, 'Middlesbrough': 13, 'West Ham': 14, 'Everton': 15, 'Southampton': 16, 'Coventry': 17, 'Sunderland': 18, \"Nott'm Forest\": 19}\n",
      "1\n",
      "./data/es-espana-master\n",
      "2017-18\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. FC Barcelona                  38  16  3  0  53:11   12  6  1  46:18    99:29  +70   93\n",
      "\n",
      "{'FC Barcelona': 0, 'Atlético Madrid': 1, 'Real Madrid CF': 2, 'Valencia CF': 3, 'Villarreal CF': 4, 'Real Betis': 5, 'Sevilla FC': 6, 'Getafe CF': 7, 'SD Eibar': 8, 'Girona FC': 9, 'Real Sociedad': 10, 'RC Celta Vigo': 11, 'RCD Español': 12, 'Deportivo Alavés': 13, 'Levante UD': 14, 'Athletic Club Bilbao': 15, 'CD Leganés': 16, 'RCD La Coruña': 17, 'UD Las Palmas': 18, 'Málaga CF': 19}\n",
      "1\n",
      "2000-01\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Real Madrid                   38  24   8   6  81:40   80\n",
      "\n",
      "{'Real Madrid': 0, 'La Coruna': 1, 'Mallorca': 2, 'Barcelona': 3, 'Valencia': 4, 'Celta': 5, 'Villarreal': 6, 'Malaga': 7, 'Espanol': 8, 'Alaves': 9, 'Las Palmas': 10, 'Vallecano': 11, 'Sociedad': 12, 'Ath Bilbao': 13, 'Zaragoza': 14, 'Valladolid': 15, 'Osasuna': 16, 'Oviedo': 17, 'Santander': 18, 'Numancia': 19}\n",
      "1\n",
      ".DS_Store\n",
      "2016-17\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. Real Madrid CF                38  14  4  1  48:20   15  2  2  58:21   106:41  +65   93\n",
      "\n",
      "{'Real Madrid CF': 0, 'FC Barcelona': 1, 'Atlético Madrid': 2, 'Sevilla FC': 3, 'Villarreal CF': 4, 'Real Sociedad': 5, 'Athletic Club Bilbao': 6, 'RCD Español': 7, 'Deportivo Alavés': 8, 'SD Eibar': 9, 'Málaga CF': 10, 'Valencia CF': 11, 'RC Celta Vigo': 12, 'UD Las Palmas': 13, 'Real Betis': 14, 'RCD La Coruña': 15, 'CD Leganés': 16, 'Sporting Gijón': 17, 'CA Osasuna': 18, 'Granada CF': 19}\n",
      "1\n",
      "SOURCES.md\n",
      "2005-06\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  25   7   6  80:35   82\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Valencia': 2, 'Sevilla': 3, 'Osasuna': 4, 'Celta': 5, 'Villarreal': 6, 'La Coruna': 7, 'Getafe': 8, 'Ath Madrid': 9, 'Zaragoza': 10, 'Ath Bilbao': 11, 'Mallorca': 12, 'Betis': 13, 'Espanol': 14, 'Santander': 15, 'Sociedad': 16, 'Alaves': 17, 'Cadiz': 18, 'Malaga': 19}\n",
      "1\n",
      "1993-94\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  25   6   7  91:42   81\n",
      "\n",
      "{'Barcelona': 0, 'La Coruna': 1, 'Zaragoza': 2, 'Real Madrid': 3, 'Ath Bilbao': 4, 'Sevilla': 5, 'Valencia': 6, 'Santander': 7, 'Tenerife': 8, 'Sp Gijon': 9, 'Oviedo': 10, 'Ath Madrid': 11, 'Sociedad': 12, 'Albacete': 13, 'Celta': 14, 'Logrones': 15, 'Vallecano': 16, 'Valladolid': 17, 'Lerida': 18, 'Osasuna': 19}\n",
      "1\n",
      "2002-03\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Real Madrid                   38  22  12   4  86:42   78\n",
      "\n",
      "{'Real Madrid': 0, 'Sociedad': 1, 'La Coruna': 2, 'Celta': 3, 'Valencia': 4, 'Barcelona': 5, 'Ath Bilbao': 6, 'Betis': 7, 'Mallorca': 8, 'Sevilla': 9, 'Ath Madrid': 10, 'Osasuna': 11, 'Valladolid': 12, 'Malaga': 13, 'Villarreal': 14, 'Santander': 15, 'Espanol': 16, 'Recreativo': 17, 'Alaves': 18, 'Vallecano': 19}\n",
      "1\n",
      "2011-12\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Real Madrid                   38  32   4   2 121:32  100\n",
      "\n",
      "{'Real Madrid': 0, 'Barcelona': 1, 'Valencia': 2, 'Malaga': 3, 'Ath Madrid': 4, 'Levante': 5, 'Osasuna': 6, 'Mallorca': 7, 'Sevilla': 8, 'Ath Bilbao': 9, 'Sociedad': 10, 'Betis': 11, 'Getafe': 12, 'Espanol': 13, 'Vallecano': 14, 'Zaragoza': 15, 'Granada': 16, 'Villarreal': 17, 'Sp Gijon': 18, 'Santander': 19}\n",
      "1\n",
      "2009-10\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  31   6   1  98:24   99\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Valencia': 2, 'Sevilla': 3, 'Mallorca': 4, 'Getafe': 5, 'Villarreal': 6, 'Ath Bilbao': 7, 'Ath Madrid': 8, 'La Coruna': 9, 'Espanol': 10, 'Osasuna': 11, 'Almeria': 12, 'Zaragoza': 13, 'Sp Gijon': 14, 'Santander': 15, 'Malaga': 16, 'Valladolid': 17, 'Tenerife': 18, 'Xerez': 19}\n",
      "1\n",
      "2003-04\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Valencia                      38  23   8   7  71:27   77\n",
      "\n",
      "{'Valencia': 0, 'Barcelona': 1, 'La Coruna': 2, 'Real Madrid': 3, 'Ath Bilbao': 4, 'Sevilla': 5, 'Ath Madrid': 6, 'Villarreal': 7, 'Betis': 8, 'Malaga': 9, 'Mallorca': 10, 'Osasuna': 11, 'Zaragoza': 12, 'Albacete': 13, 'Sociedad': 14, 'Santander': 15, 'Espanol': 16, 'Valladolid': 17, 'Celta': 18, 'Murcia': 19}\n",
      "1\n",
      "2014-15\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. FC Barcelona                  38  16  1  2  64:11   14  3  2  46:10   110:21  +89   94\n",
      "\n",
      "{'FC Barcelona': 0, 'Real Madrid CF': 1, 'Atlético Madrid': 2, 'Valencia CF': 3, 'Sevilla FC': 4, 'Villarreal CF': 5, 'Athletic Club Bilbao': 6, 'RC Celta Vigo': 7, 'Málaga CF': 8, 'RCD Español': 9, 'Rayo Vallecano': 10, 'Real Sociedad': 11, 'Elche CF': 12, 'Getafe CF': 13, 'Levante UD': 14, 'SD Eibar': 15, 'RCD La Coruña': 16, 'Granada CF': 17, 'UD Almería': 18, 'Córdoba CF': 19}\n",
      "1\n",
      "README.md\n",
      "2004-05\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  25   9   4  73:29   84\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Villarreal': 2, 'Betis': 3, 'Espanol': 4, 'Sevilla': 5, 'Valencia': 6, 'Ath Bilbao': 7, 'La Coruna': 8, 'Malaga': 9, 'Ath Madrid': 10, 'Zaragoza': 11, 'Getafe': 12, 'Sociedad': 13, 'Osasuna': 14, 'Santander': 15, 'Mallorca': 16, 'Levante': 17, 'Numancia': 18, 'Albacete': 19}\n",
      "1\n",
      "1999-00\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. La Coruna                     38  21   6  11  66:44   69\n",
      "\n",
      "{'La Coruna': 0, 'Barcelona': 1, 'Valencia': 2, 'Zaragoza': 3, 'Real Madrid': 4, 'Alaves': 5, 'Celta': 6, 'Valladolid': 7, 'Vallecano': 8, 'Mallorca': 9, 'Ath Bilbao': 10, 'Malaga': 11, 'Espanol': 12, 'Sociedad': 13, 'Santander': 14, 'Numancia': 15, 'Oviedo': 16, 'Betis': 17, 'Ath Madrid': 18, 'Sevilla': 19}\n",
      "1\n",
      "2008-09\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  27   6   5 105:35   87\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Sevilla': 2, 'Ath Madrid': 3, 'Villarreal': 4, 'Valencia': 5, 'La Coruna': 6, 'Malaga': 7, 'Mallorca': 8, 'Espanol': 9, 'Santander': 10, 'Almeria': 11, 'Ath Bilbao': 12, 'Osasuna': 13, 'Valladolid': 14, 'Sp Gijon': 15, 'Getafe': 16, 'Betis': 17, 'Numancia': 18, 'Recreativo': 19}\n",
      "1\n",
      "2013-14\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Ath Madrid                    38  28   6   4  77:26   90\n",
      "\n",
      "{'Ath Madrid': 0, 'Barcelona': 1, 'Real Madrid': 2, 'Ath Bilbao': 3, 'Sevilla': 4, 'Villarreal': 5, 'Sociedad': 6, 'Valencia': 7, 'Celta': 8, 'Levante': 9, 'Malaga': 10, 'Vallecano': 11, 'Espanol': 12, 'Getafe': 13, 'Granada': 14, 'Elche': 15, 'Almeria': 16, 'Osasuna': 17, 'Valladolid': 18, 'Betis': 19}\n",
      "1\n",
      "1994-95\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Real Madrid                   38  23   9   6  76:29   78\n",
      "\n",
      "{'Real Madrid': 0, 'La Coruna': 1, 'Barcelona': 2, 'Betis': 3, 'Zaragoza': 4, 'Sevilla': 5, 'Ath Bilbao': 6, 'Espanol': 7, 'Oviedo': 8, 'Valencia': 9, 'Sociedad': 10, 'Santander': 11, 'Ath Madrid': 12, 'Tenerife': 13, 'Celta': 14, 'Compostela': 15, 'Albacete': 16, 'Sp Gijon': 17, 'Valladolid': 18, 'Logrones': 19}\n",
      "1\n",
      "2001-02\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Valencia                      38  21  12   5  51:27   75\n",
      "\n",
      "{'Valencia': 0, 'La Coruna': 1, 'Real Madrid': 2, 'Barcelona': 3, 'Celta': 4, 'Betis': 5, 'Alaves': 6, 'Sevilla': 7, 'Malaga': 8, 'Ath Bilbao': 9, 'Vallecano': 10, 'Valladolid': 11, 'Sociedad': 12, 'Espanol': 13, 'Villarreal': 14, 'Mallorca': 15, 'Osasuna': 16, 'Las Palmas': 17, 'Tenerife': 18, 'Zaragoza': 19}\n",
      "1\n",
      "2012-13\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  32   4   2 115:40  100\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Ath Madrid': 2, 'Sociedad': 3, 'Valencia': 4, 'Malaga': 5, 'Betis': 6, 'Vallecano': 7, 'Sevilla': 8, 'Getafe': 9, 'Levante': 10, 'Ath Bilbao': 11, 'Espanol': 12, 'Valladolid': 13, 'Granada': 14, 'Osasuna': 15, 'Celta': 16, 'Mallorca': 17, 'La Coruna': 18, 'Zaragoza': 19}\n",
      "1\n",
      "1998-99\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  24   7   7  87:43   79\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Mallorca': 2, 'Valencia': 3, 'Celta': 4, 'La Coruna': 5, 'Espanol': 6, 'Ath Bilbao': 7, 'Zaragoza': 8, 'Sociedad': 9, 'Betis': 10, 'Valladolid': 11, 'Ath Madrid': 12, 'Oviedo': 13, 'Santander': 14, 'Alaves': 15, 'Extremadura': 16, 'Villareal': 17, 'Tenerife': 18, 'Salamanca': 19}\n",
      "1\n",
      "2015-16\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "```\n",
      "\n",
      "                                        - Home -          - Away -            - Total -\n",
      "\n",
      "                                 Pld   W  D  L   F:A     W  D  L   F:A      F:A   +/-  Pts\n",
      "\n",
      " 1. FC Barcelona                  38  16  1  2  67:14   13  3  3  45:15   112:29  +83   91\n",
      "\n",
      "{'FC Barcelona': 0, 'Real Madrid CF': 1, 'Atlético Madrid': 2, 'Villarreal CF': 3, 'Athletic Club Bilbao': 4, 'RC Celta Vigo': 5, 'Sevilla FC': 6, 'Málaga CF': 7, 'Real Sociedad': 8, 'Real Betis': 9, 'Valencia CF': 10, 'UD Las Palmas': 11, 'SD Eibar': 12, 'RCD Español': 13, 'RCD La Coruña': 14, 'Sporting Gijón': 15, 'Granada CF': 16, 'Rayo Vallecano': 17, 'Getafe CF': 18, 'Levante UD': 19}\n",
      "1\n",
      "2006-07\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  22  10   6  78:33   76\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Sevilla': 2, 'Valencia': 3, 'Villarreal': 4, 'Zaragoza': 5, 'Ath Madrid': 6, 'Recreativo': 7, 'Getafe': 8, 'Santander': 9, 'Mallorca': 10, 'Espanol': 11, 'La Coruna': 12, 'Osasuna': 13, 'Levante': 14, 'Betis': 15, 'Ath Bilbao': 16, 'Celta': 17, 'Sociedad': 18, 'Gimnastic': 19}\n",
      "1\n",
      "1997-98\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  23   5  10  78:56   74\n",
      "\n",
      "{'Barcelona': 0, 'Ath Bilbao': 1, 'Sociedad': 2, 'Real Madrid': 3, 'Ath Madrid': 4, 'Mallorca': 5, 'Celta': 6, 'Betis': 7, 'Valencia': 8, 'Espanol': 9, 'Valladolid': 10, 'La Coruna': 11, 'Zaragoza': 12, 'Salamanca': 13, 'Santander': 14, 'Tenerife': 15, 'Compostela': 16, 'Oviedo': 17, 'Merida': 18, 'Sp Gijon': 19}\n",
      "1\n",
      "2010-11\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Barcelona                     38  30   6   2  95:21   96\n",
      "\n",
      "{'Barcelona': 0, 'Real Madrid': 1, 'Valencia': 2, 'Villarreal': 3, 'Ath Madrid': 4, 'Ath Bilbao': 5, 'Sevilla': 6, 'Espanol': 7, 'Osasuna': 8, 'Sp Gijon': 9, 'Malaga': 10, 'Santander': 11, 'Levante': 12, 'Zaragoza': 13, 'Sociedad': 14, 'Getafe': 15, 'Mallorca': 16, 'La Coruna': 17, 'Hercules': 18, 'Almeria': 19}\n",
      "1\n",
      "2007-08\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Real Madrid                   38  27   4   7  84:36   85\n",
      "\n",
      "{'Real Madrid': 0, 'Villarreal': 1, 'Barcelona': 2, 'Sevilla': 3, 'Ath Madrid': 4, 'Santander': 5, 'Mallorca': 6, 'La Coruna': 7, 'Almeria': 8, 'Valencia': 9, 'Ath Bilbao': 10, 'Espanol': 11, 'Getafe': 12, 'Betis': 13, 'Valladolid': 14, 'Recreativo': 15, 'Osasuna': 16, 'Zaragoza': 17, 'Murcia': 18, 'Levante': 19}\n",
      "1\n",
      "./data/it-italy-master\n",
      ".DS_Store\n",
      "SOURCES.md\n",
      "2005-06\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Juventus                      38  27  10   1  71:24   91\n",
      "\n",
      "{'Juventus': 0, 'Milan': 1, 'Inter': 2, 'Fiorentina': 3, 'Roma': 4, 'Lazio': 5, 'Chievo': 6, 'Palermo': 7, 'Livorno': 8, 'Parma': 9, 'Empoli': 10, 'Ascoli': 11, 'Udinese': 12, 'Sampdoria': 13, 'Reggina': 14, 'Cagliari': 15, 'Siena': 16, 'Messina': 17, 'Lecce': 18, 'Treviso': 19}\n",
      "1\n",
      "2011-12\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Juventus                      38  23  15   0  68:20   84\n",
      "\n",
      "{'Juventus': 0, 'Milan': 1, 'Udinese': 2, 'Lazio': 3, 'Napoli': 4, 'Inter': 5, 'Roma': 6, 'Parma': 7, 'Atalanta': 8, 'Bologna': 9, 'Chievo': 10, 'Catania': 11, 'Fiorentina': 12, 'Siena': 13, 'Cagliari': 14, 'Palermo': 15, 'Genoa': 16, 'Lecce': 17, 'Novara': 18, 'Cesena': 19}\n",
      "1\n",
      "2009-10\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Inter                         38  24  10   4  75:34   82\n",
      "\n",
      "{'Inter': 0, 'Roma': 1, 'Milan': 2, 'Sampdoria': 3, 'Palermo': 4, 'Napoli': 5, 'Juventus': 6, 'Parma': 7, 'Genoa': 8, 'Bari': 9, 'Fiorentina': 10, 'Lazio': 11, 'Catania': 12, 'Cagliari': 13, 'Udinese': 14, 'Chievo': 15, 'Bologna': 16, 'Atalanta': 17, 'Siena': 18, 'Livorno': 19}\n",
      "1\n",
      "README.md\n",
      "2004-05\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Juventus                      38  26   8   4  67:27   86\n",
      "\n",
      "{'Juventus': 0, 'Milan': 1, 'Inter': 2, 'Udinese': 3, 'Sampdoria': 4, 'Palermo': 5, 'Messina': 6, 'Roma': 7, 'Livorno': 8, 'Lazio': 9, 'Lecce': 10, 'Cagliari': 11, 'Reggina': 12, 'Siena': 13, 'Chievo': 14, 'Bologna': 15, 'Fiorentina': 16, 'Parma': 17, 'Brescia': 18, 'Atalanta': 19}\n",
      "1\n",
      "2008-09\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Inter                         38  25   9   4  70:32   84\n",
      "\n",
      "{'Inter': 0, 'Milan': 1, 'Juventus': 2, 'Genoa': 3, 'Fiorentina': 4, 'Roma': 5, 'Udinese': 6, 'Palermo': 7, 'Cagliari': 8, 'Lazio': 9, 'Atalanta': 10, 'Napoli': 11, 'Sampdoria': 12, 'Siena': 13, 'Catania': 14, 'Chievo': 15, 'Bologna': 16, 'Torino': 17, 'Reggina': 18, 'Lecce': 19}\n",
      "1\n",
      "2013-14\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Juventus                      38  33   3   2  80:23  102\n",
      "\n",
      "{'Juventus': 0, 'Roma': 1, 'Napoli': 2, 'Fiorentina': 3, 'Inter': 4, 'Parma': 5, 'Torino': 6, 'Milan': 7, 'Lazio': 8, 'Verona': 9, 'Atalanta': 10, 'Sampdoria': 11, 'Genoa': 12, 'Udinese': 13, 'Cagliari': 14, 'Chievo': 15, 'Sassuolo': 16, 'Catania': 17, 'Bologna': 18, 'Livorno': 19}\n",
      "1\n",
      "2012-13\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Juventus                      38  27   6   5  71:24   87\n",
      "\n",
      "{'Juventus': 0, 'Napoli': 1, 'Milan': 2, 'Fiorentina': 3, 'Udinese': 4, 'Roma': 5, 'Lazio': 6, 'Catania': 7, 'Inter': 8, 'Parma': 9, 'Cagliari': 10, 'Chievo': 11, 'Bologna': 12, 'Sampdoria': 13, 'Atalanta': 14, 'Torino': 15, 'Genoa': 16, 'Siena': 17, 'Palermo': 18, 'Pescara': 19}\n",
      "1\n",
      "2006-07\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Inter                         38  30   7   1  80:34   97\n",
      "\n",
      "{'Inter': 0, 'Roma': 1, 'Fiorentina': 2, 'Milan': 3, 'Lazio': 4, 'Palermo': 5, 'Empoli': 6, 'Reggina': 7, 'Atalanta': 8, 'Sampdoria': 9, 'Udinese': 10, 'Livorno': 11, 'Parma': 12, 'Siena': 13, 'Catania': 14, 'Cagliari': 15, 'Torino': 16, 'Chievo': 17, 'Ascoli': 18, 'Messina': 19}\n",
      "1\n",
      "2010-11\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Milan                         38  24  10   4  65:24   82\n",
      "\n",
      "{'Milan': 0, 'Inter': 1, 'Napoli': 2, 'Udinese': 3, 'Lazio': 4, 'Roma': 5, 'Juventus': 6, 'Palermo': 7, 'Fiorentina': 8, 'Genoa': 9, 'Chievo': 10, 'Parma': 11, 'Catania': 12, 'Cagliari': 13, 'Bologna': 14, 'Cesena': 15, 'Lecce': 16, 'Sampdoria': 17, 'Brescia': 18, 'Bari': 19}\n",
      "1\n",
      "2007-08\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "### Standings\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "~~~\n",
      "\n",
      " 1. Inter                         38  25  10   3  69:26   85\n",
      "\n",
      "{'Inter': 0, 'Roma': 1, 'Juventus': 2, 'Fiorentina': 3, 'Milan': 4, 'Sampdoria': 5, 'Udinese': 6, 'Napoli': 7, 'Atalanta': 8, 'Genoa': 9, 'Palermo': 10, 'Lazio': 11, 'Siena': 12, 'Cagliari': 13, 'Torino': 14, 'Reggina': 15, 'Catania': 16, 'Empoli': 17, 'Parma': 18, 'Livorno': 19}\n",
      "1\n"
     ]
    }
   ],
   "source": [
    "country_dirs = []\n",
    "season_dirs = []\n",
    "\n",
    "nseasons = 0\n",
    "team_maps = []\n",
    "game_mats = []\n",
    "\n",
    "for entry in os.scandir('./data'):\n",
    "    if entry.is_dir():\n",
    "        country_dirs.append(entry.path)\n",
    "\n",
    "for cdir in country_dirs:\n",
    "#     print(cdir)\n",
    "    for entry in os.scandir(cdir):\n",
    "#         print(entry.name)\n",
    "        if entry.is_dir():\n",
    "            season_dir = entry.path\n",
    "            team_maps.append(season_team_map(season_dir))\n",
    "            game_mats.append(season_game_mat(season_dir, team_maps[nseasons]))\n",
    "            nseasons += 1\n",
    "            if(cdir=='./data/fr-france-master' and int(entry.name.split('-')[0])>=2002):\n",
    "                team_maps.append(season_team_map_div_2(season_dir))\n",
    "                game_mats.append(season_game_mat(season_dir, team_maps[nseasons], div=2))\n",
    "                nseasons += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "88"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nseasons"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def eucl_error(outputs, targets):\n",
    "    return np.mean(np.abs(targets - outputs))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def disc_error(outputs, targets):\n",
    "    rounded_outputs = np.round(outputs)\n",
    "    err = 0\n",
    "    n = 0\n",
    "    for i in range(len(outputs)):\n",
    "        for j in range(len(outputs[i])):\n",
    "            if(int(rounded_outputs[i][j]) != int(targets[i][j])):\n",
    "                err += 1\n",
    "            n += 1\n",
    "    return float(err)/n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ntraining = int(nseasons*0.7)\n",
    "ntesting = nseasons - ntraining\n",
    "\n",
    "training_game_mats = game_mats[:ntraining]\n",
    "testing_game_mats = game_mats[ntesting:]\n",
    "\n",
    "X_training = np.reshape(np.array(training_game_mats), (ntraining*nmatches, nteams))\n",
    "lengths = np.full(ntraining, nmatches)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_trainings = []\n",
    "Y_trainings = []\n",
    "for i in range(1, nmatches):\n",
    "    X_training = []\n",
    "    Y_training = []\n",
    "    for game_mat in training_game_mats:\n",
    "        for j in range(nmatches-i):\n",
    "            X_training.append(np.ndarray.flatten(game_mat[j:j+i]))\n",
    "            Y_training.append(game_mat[j+i])\n",
    "    \n",
    "    X_training = np.array(X_training).reshape(-1,i,nteams)\n",
    "    Y_training = np.array(Y_training).reshape(-1,nteams)\n",
    "    X_trainings.append(X_training)\n",
    "    Y_trainings.append(Y_training)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_testings = []\n",
    "Y_testings = []\n",
    "for i in range(1, nmatches):\n",
    "    X_testing = []\n",
    "    Y_testing = []\n",
    "    for game_mat in testing_game_mats:\n",
    "        for j in range(nmatches-i):\n",
    "            X_testing.append(np.ndarray.flatten(game_mat[j:j+i]))\n",
    "            Y_testing.append(game_mat[j+i])\n",
    "            \n",
    "    X_testing = np.array(X_testing).reshape(-1,i,nteams)\n",
    "    Y_testing = np.array(Y_testing).reshape(-1,nteams)\n",
    "    X_testings.append(X_testing)\n",
    "    Y_testings.append(Y_testing)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# layer_sizes = [30, 50, 100, 200]\n",
    "layer_sizes = [200]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rnns = {}\n",
    "for size in layer_sizes:\n",
    "    key = str(size)+'_neuron_layers'\n",
    "    rnns[key] = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'200_neuron_layers': []}"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "histories = {}\n",
    "for size in layer_sizes:\n",
    "    key = str(size)+'_neuron_layers'\n",
    "    histories[key] = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "epochs = 20\n",
    "max_sequence_length = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sequence Length of 0\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "2257/2257 [==============================] - 4s 2ms/step - loss: 0.6878\n",
      "Epoch 2/20\n",
      "2257/2257 [==============================] - 1s 268us/step - loss: 0.6645\n",
      "Epoch 3/20\n",
      "2257/2257 [==============================] - 1s 241us/step - loss: 0.6601\n",
      "Epoch 4/20\n",
      "2257/2257 [==============================] - 1s 246us/step - loss: 0.6575\n",
      "Epoch 5/20\n",
      "2257/2257 [==============================] - 1s 249us/step - loss: 0.6558\n",
      "Epoch 6/20\n",
      "2257/2257 [==============================] - 1s 257us/step - loss: 0.6547\n",
      "Epoch 7/20\n",
      "2257/2257 [==============================] - 1s 281us/step - loss: 0.6536\n",
      "Epoch 8/20\n",
      "2257/2257 [==============================] - 1s 288us/step - loss: 0.6529\n",
      "Epoch 9/20\n",
      "2257/2257 [==============================] - 1s 269us/step - loss: 0.6525\n",
      "Epoch 10/20\n",
      "2257/2257 [==============================] - 1s 258us/step - loss: 0.6513\n",
      "Epoch 11/20\n",
      "2257/2257 [==============================] - 1s 259us/step - loss: 0.6504\n",
      "Epoch 12/20\n",
      "2257/2257 [==============================] - 1s 263us/step - loss: 0.6493\n",
      "Epoch 13/20\n",
      "2257/2257 [==============================] - 1s 318us/step - loss: 0.6482\n",
      "Epoch 14/20\n",
      "2257/2257 [==============================] - 1s 330us/step - loss: 0.6477\n",
      "Epoch 15/20\n",
      "2257/2257 [==============================] - 1s 277us/step - loss: 0.6459\n",
      "Epoch 16/20\n",
      "2257/2257 [==============================] - 1s 268us/step - loss: 0.6449\n",
      "Epoch 17/20\n",
      "2257/2257 [==============================] - 1s 262us/step - loss: 0.6439\n",
      "Epoch 18/20\n",
      "2257/2257 [==============================] - 1s 279us/step - loss: 0.6424\n",
      "Epoch 19/20\n",
      "2257/2257 [==============================] - 1s 293us/step - loss: 0.6410\n",
      "Epoch 20/20\n",
      "2257/2257 [==============================] - 1s 260us/step - loss: 0.6395\n",
      "Sequence Length of 1\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "2196/2196 [==============================] - 4s 2ms/step - loss: 0.6812\n",
      "Epoch 2/20\n",
      "2196/2196 [==============================] - 1s 314us/step - loss: 0.6630\n",
      "Epoch 3/20\n",
      "2196/2196 [==============================] - 1s 312us/step - loss: 0.6584\n",
      "Epoch 4/20\n",
      "2196/2196 [==============================] - 1s 314us/step - loss: 0.6547\n",
      "Epoch 5/20\n",
      "2196/2196 [==============================] - 1s 350us/step - loss: 0.6517\n",
      "Epoch 6/20\n",
      "2196/2196 [==============================] - 1s 326us/step - loss: 0.6502\n",
      "Epoch 7/20\n",
      "2196/2196 [==============================] - 1s 372us/step - loss: 0.6474\n",
      "Epoch 8/20\n",
      "2196/2196 [==============================] - 1s 317us/step - loss: 0.6448\n",
      "Epoch 9/20\n",
      "2196/2196 [==============================] - 1s 343us/step - loss: 0.6412\n",
      "Epoch 10/20\n",
      "2196/2196 [==============================] - 1s 356us/step - loss: 0.6394\n",
      "Epoch 11/20\n",
      "2196/2196 [==============================] - 1s 416us/step - loss: 0.6349\n",
      "Epoch 12/20\n",
      "2196/2196 [==============================] - 1s 384us/step - loss: 0.6312\n",
      "Epoch 13/20\n",
      "2196/2196 [==============================] - 1s 412us/step - loss: 0.6261\n",
      "Epoch 14/20\n",
      "2196/2196 [==============================] - 1s 321us/step - loss: 0.6208\n",
      "Epoch 15/20\n",
      "2196/2196 [==============================] - 1s 337us/step - loss: 0.6131\n",
      "Epoch 16/20\n",
      "2196/2196 [==============================] - 1s 346us/step - loss: 0.6065\n",
      "Epoch 17/20\n",
      "2196/2196 [==============================] - 1s 322us/step - loss: 0.5978\n",
      "Epoch 18/20\n",
      "2196/2196 [==============================] - 1s 324us/step - loss: 0.5889\n",
      "Epoch 19/20\n",
      "2196/2196 [==============================] - 1s 347us/step - loss: 0.5795\n",
      "Epoch 20/20\n",
      "2196/2196 [==============================] - 1s 365us/step - loss: 0.5701\n",
      "Sequence Length of 2\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "2135/2135 [==============================] - 4s 2ms/step - loss: 0.6791\n",
      "Epoch 2/20\n",
      "2135/2135 [==============================] - 1s 371us/step - loss: 0.6628\n",
      "Epoch 3/20\n",
      "2135/2135 [==============================] - 1s 379us/step - loss: 0.6581\n",
      "Epoch 4/20\n",
      "2135/2135 [==============================] - 1s 391us/step - loss: 0.6552\n",
      "Epoch 5/20\n",
      "2135/2135 [==============================] - 1s 428us/step - loss: 0.6529\n",
      "Epoch 6/20\n",
      "2135/2135 [==============================] - 1s 378us/step - loss: 0.6508\n",
      "Epoch 7/20\n",
      "2135/2135 [==============================] - 1s 526us/step - loss: 0.6471\n",
      "Epoch 8/20\n",
      "2135/2135 [==============================] - 1s 491us/step - loss: 0.6450\n",
      "Epoch 9/20\n",
      "2135/2135 [==============================] - 1s 408us/step - loss: 0.6419\n",
      "Epoch 10/20\n",
      "2135/2135 [==============================] - 1s 400us/step - loss: 0.6387\n",
      "Epoch 11/20\n",
      "2135/2135 [==============================] - 1s 402us/step - loss: 0.6345\n",
      "Epoch 12/20\n",
      "2135/2135 [==============================] - 1s 396us/step - loss: 0.6288\n",
      "Epoch 13/20\n",
      "2135/2135 [==============================] - 1s 402us/step - loss: 0.6214\n",
      "Epoch 14/20\n",
      "2135/2135 [==============================] - 1s 410us/step - loss: 0.6141\n",
      "Epoch 15/20\n",
      "2135/2135 [==============================] - 1s 403us/step - loss: 0.6038\n",
      "Epoch 16/20\n",
      "2135/2135 [==============================] - 1s 396us/step - loss: 0.5919\n",
      "Epoch 17/20\n",
      "2135/2135 [==============================] - 1s 401us/step - loss: 0.5805\n",
      "Epoch 18/20\n",
      "2135/2135 [==============================] - 1s 418us/step - loss: 0.5646\n",
      "Epoch 19/20\n",
      "2135/2135 [==============================] - 1s 467us/step - loss: 0.5508\n",
      "Epoch 20/20\n",
      "2135/2135 [==============================] - 1s 433us/step - loss: 0.5370\n",
      "Sequence Length of 3\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "2074/2074 [==============================] - 5s 2ms/step - loss: 0.6774\n",
      "Epoch 2/20\n",
      "2074/2074 [==============================] - 1s 441us/step - loss: 0.6620\n",
      "Epoch 3/20\n",
      "2074/2074 [==============================] - 1s 449us/step - loss: 0.6568\n",
      "Epoch 4/20\n",
      "2074/2074 [==============================] - 1s 438us/step - loss: 0.6536\n",
      "Epoch 5/20\n",
      "2074/2074 [==============================] - 1s 469us/step - loss: 0.6499\n",
      "Epoch 6/20\n",
      "2074/2074 [==============================] - 1s 513us/step - loss: 0.6479\n",
      "Epoch 7/20\n",
      "2074/2074 [==============================] - 1s 568us/step - loss: 0.6458\n",
      "Epoch 8/20\n",
      "2074/2074 [==============================] - 1s 449us/step - loss: 0.6417\n",
      "Epoch 9/20\n",
      "2074/2074 [==============================] - 1s 440us/step - loss: 0.6387\n",
      "Epoch 10/20\n",
      "2074/2074 [==============================] - 1s 444us/step - loss: 0.6343\n",
      "Epoch 11/20\n",
      "2074/2074 [==============================] - 1s 454us/step - loss: 0.6286\n",
      "Epoch 12/20\n",
      "2074/2074 [==============================] - 1s 474us/step - loss: 0.6230\n",
      "Epoch 13/20\n",
      "2074/2074 [==============================] - 1s 500us/step - loss: 0.6142\n",
      "Epoch 14/20\n",
      "2074/2074 [==============================] - 1s 539us/step - loss: 0.6033\n",
      "Epoch 15/20\n",
      "2074/2074 [==============================] - 1s 551us/step - loss: 0.5908\n",
      "Epoch 16/20\n",
      "2074/2074 [==============================] - 1s 630us/step - loss: 0.5775\n",
      "Epoch 17/20\n",
      "2074/2074 [==============================] - 1s 669us/step - loss: 0.5608\n",
      "Epoch 18/20\n",
      "2074/2074 [==============================] - 1s 613us/step - loss: 0.5444\n",
      "Epoch 19/20\n",
      "2074/2074 [==============================] - 1s 579us/step - loss: 0.5293\n",
      "Epoch 20/20\n",
      "2074/2074 [==============================] - 1s 587us/step - loss: 0.5126\n",
      "Sequence Length of 4\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "2013/2013 [==============================] - 6s 3ms/step - loss: 0.6768\n",
      "Epoch 2/20\n",
      "2013/2013 [==============================] - 1s 572us/step - loss: 0.6614 0s - lo\n",
      "Epoch 3/20\n",
      "2013/2013 [==============================] - 1s 556us/step - loss: 0.6562\n",
      "Epoch 4/20\n",
      "2013/2013 [==============================] - 1s 669us/step - loss: 0.6532\n",
      "Epoch 5/20\n",
      "2013/2013 [==============================] - 1s 619us/step - loss: 0.6503\n",
      "Epoch 6/20\n",
      "2013/2013 [==============================] - 1s 662us/step - loss: 0.6481\n",
      "Epoch 7/20\n",
      "2013/2013 [==============================] - 1s 622us/step - loss: 0.6444\n",
      "Epoch 8/20\n",
      "2013/2013 [==============================] - 2s 761us/step - loss: 0.6414\n",
      "Epoch 9/20\n",
      "2013/2013 [==============================] - 1s 674us/step - loss: 0.6381\n",
      "Epoch 10/20\n",
      "2013/2013 [==============================] - 1s 598us/step - loss: 0.6337\n",
      "Epoch 11/20\n",
      "2013/2013 [==============================] - 1s 573us/step - loss: 0.6284\n",
      "Epoch 12/20\n",
      "2013/2013 [==============================] - 1s 611us/step - loss: 0.6212\n",
      "Epoch 13/20\n",
      "2013/2013 [==============================] - 1s 587us/step - loss: 0.6103\n",
      "Epoch 14/20\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2013/2013 [==============================] - 1s 475us/step - loss: 0.6002\n",
      "Epoch 15/20\n",
      "2013/2013 [==============================] - 1s 461us/step - loss: 0.5871\n",
      "Epoch 16/20\n",
      "2013/2013 [==============================] - 1s 461us/step - loss: 0.5724\n",
      "Epoch 17/20\n",
      "2013/2013 [==============================] - 1s 592us/step - loss: 0.5539\n",
      "Epoch 18/20\n",
      "2013/2013 [==============================] - 1s 450us/step - loss: 0.5381\n",
      "Epoch 19/20\n",
      "2013/2013 [==============================] - 1s 435us/step - loss: 0.5206\n",
      "Epoch 20/20\n",
      "2013/2013 [==============================] - 1s 440us/step - loss: 0.4957\n",
      "Sequence Length of 5\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "1952/1952 [==============================] - 5s 3ms/step - loss: 0.6757\n",
      "Epoch 2/20\n",
      "1952/1952 [==============================] - 1s 495us/step - loss: 0.6601\n",
      "Epoch 3/20\n",
      "1952/1952 [==============================] - 1s 440us/step - loss: 0.6540\n",
      "Epoch 4/20\n",
      "1952/1952 [==============================] - 1s 443us/step - loss: 0.6516\n",
      "Epoch 5/20\n",
      "1952/1952 [==============================] - 1s 448us/step - loss: 0.6486\n",
      "Epoch 6/20\n",
      "1952/1952 [==============================] - 1s 465us/step - loss: 0.6457\n",
      "Epoch 7/20\n",
      "1952/1952 [==============================] - 1s 560us/step - loss: 0.6419\n",
      "Epoch 8/20\n",
      "1952/1952 [==============================] - 1s 495us/step - loss: 0.6397\n",
      "Epoch 9/20\n",
      "1952/1952 [==============================] - 1s 471us/step - loss: 0.6353\n",
      "Epoch 10/20\n",
      "1952/1952 [==============================] - 1s 477us/step - loss: 0.6304\n",
      "Epoch 11/20\n",
      "1952/1952 [==============================] - 1s 469us/step - loss: 0.6245\n",
      "Epoch 12/20\n",
      "1952/1952 [==============================] - 1s 450us/step - loss: 0.6171\n",
      "Epoch 13/20\n",
      "1952/1952 [==============================] - 1s 516us/step - loss: 0.6067\n",
      "Epoch 14/20\n",
      "1952/1952 [==============================] - 1s 488us/step - loss: 0.5940\n",
      "Epoch 15/20\n",
      "1952/1952 [==============================] - 1s 450us/step - loss: 0.5815\n",
      "Epoch 16/20\n",
      "1952/1952 [==============================] - 1s 455us/step - loss: 0.5657\n",
      "Epoch 17/20\n",
      "1952/1952 [==============================] - 1s 439us/step - loss: 0.5469\n",
      "Epoch 18/20\n",
      "1952/1952 [==============================] - 1s 446us/step - loss: 0.5266\n",
      "Epoch 19/20\n",
      "1952/1952 [==============================] - 1s 447us/step - loss: 0.5053\n",
      "Epoch 20/20\n",
      "1952/1952 [==============================] - 1s 444us/step - loss: 0.4846\n",
      "Sequence Length of 6\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "1891/1891 [==============================] - 5s 3ms/step - loss: 0.6787\n",
      "Epoch 2/20\n",
      "1891/1891 [==============================] - 1s 467us/step - loss: 0.6608\n",
      "Epoch 3/20\n",
      "1891/1891 [==============================] - 1s 463us/step - loss: 0.6551\n",
      "Epoch 4/20\n",
      "1891/1891 [==============================] - 1s 461us/step - loss: 0.6516\n",
      "Epoch 5/20\n",
      "1891/1891 [==============================] - 1s 454us/step - loss: 0.6492\n",
      "Epoch 6/20\n",
      "1891/1891 [==============================] - 1s 465us/step - loss: 0.6463\n",
      "Epoch 7/20\n",
      "1891/1891 [==============================] - 1s 463us/step - loss: 0.6441\n",
      "Epoch 8/20\n",
      "1891/1891 [==============================] - 1s 480us/step - loss: 0.6412\n",
      "Epoch 9/20\n",
      "1891/1891 [==============================] - 1s 476us/step - loss: 0.6360\n",
      "Epoch 10/20\n",
      "1891/1891 [==============================] - 1s 501us/step - loss: 0.6332 0s -\n",
      "Epoch 11/20\n",
      "1891/1891 [==============================] - 1s 527us/step - loss: 0.6259\n",
      "Epoch 12/20\n",
      "1891/1891 [==============================] - 1s 492us/step - loss: 0.6208\n",
      "Epoch 13/20\n",
      "1891/1891 [==============================] - 1s 498us/step - loss: 0.6117\n",
      "Epoch 14/20\n",
      "1891/1891 [==============================] - 1s 492us/step - loss: 0.6017\n",
      "Epoch 15/20\n",
      "1891/1891 [==============================] - 1s 486us/step - loss: 0.5918\n",
      "Epoch 16/20\n",
      "1891/1891 [==============================] - 1s 522us/step - loss: 0.5781\n",
      "Epoch 17/20\n",
      "1891/1891 [==============================] - 1s 569us/step - loss: 0.5600\n",
      "Epoch 18/20\n",
      "1891/1891 [==============================] - 1s 502us/step - loss: 0.5467\n",
      "Epoch 19/20\n",
      "1891/1891 [==============================] - 1s 481us/step - loss: 0.5295\n",
      "Epoch 20/20\n",
      "1891/1891 [==============================] - 1s 480us/step - loss: 0.5084\n",
      "Sequence Length of 7\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "1830/1830 [==============================] - 5s 3ms/step - loss: 0.6773\n",
      "Epoch 2/20\n",
      "1830/1830 [==============================] - 1s 484us/step - loss: 0.6619\n",
      "Epoch 3/20\n",
      "1830/1830 [==============================] - 1s 486us/step - loss: 0.6561\n",
      "Epoch 4/20\n",
      "1830/1830 [==============================] - 1s 493us/step - loss: 0.6517\n",
      "Epoch 5/20\n",
      "1830/1830 [==============================] - 1s 496us/step - loss: 0.6483\n",
      "Epoch 6/20\n",
      "1830/1830 [==============================] - 1s 501us/step - loss: 0.6450\n",
      "Epoch 7/20\n",
      "1830/1830 [==============================] - 1s 574us/step - loss: 0.6413\n",
      "Epoch 8/20\n",
      "1830/1830 [==============================] - 1s 509us/step - loss: 0.6377\n",
      "Epoch 9/20\n",
      "1830/1830 [==============================] - 1s 532us/step - loss: 0.6345\n",
      "Epoch 10/20\n",
      "1830/1830 [==============================] - 1s 631us/step - loss: 0.6296\n",
      "Epoch 11/20\n",
      "1830/1830 [==============================] - 1s 532us/step - loss: 0.6221\n",
      "Epoch 12/20\n",
      "1830/1830 [==============================] - 1s 542us/step - loss: 0.6138\n",
      "Epoch 13/20\n",
      "1830/1830 [==============================] - 1s 560us/step - loss: 0.6054\n",
      "Epoch 14/20\n",
      "1830/1830 [==============================] - 1s 587us/step - loss: 0.5934\n",
      "Epoch 15/20\n",
      "1830/1830 [==============================] - 1s 552us/step - loss: 0.5802\n",
      "Epoch 16/20\n",
      "1830/1830 [==============================] - 1s 569us/step - loss: 0.5627\n",
      "Epoch 17/20\n",
      "1830/1830 [==============================] - 1s 601us/step - loss: 0.5443\n",
      "Epoch 18/20\n",
      "1830/1830 [==============================] - 1s 543us/step - loss: 0.5277\n",
      "Epoch 19/20\n",
      "1830/1830 [==============================] - 1s 545us/step - loss: 0.5081\n",
      "Epoch 20/20\n",
      "1830/1830 [==============================] - 1s 546us/step - loss: 0.4857\n",
      "Sequence Length of 8\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "1769/1769 [==============================] - 5s 3ms/step - loss: 0.6778\n",
      "Epoch 2/20\n",
      "1769/1769 [==============================] - 1s 587us/step - loss: 0.6598\n",
      "Epoch 3/20\n",
      "1769/1769 [==============================] - 1s 541us/step - loss: 0.6553\n",
      "Epoch 4/20\n",
      "1769/1769 [==============================] - 1s 541us/step - loss: 0.6516\n",
      "Epoch 5/20\n",
      "1769/1769 [==============================] - 1s 557us/step - loss: 0.6478\n",
      "Epoch 6/20\n",
      "1769/1769 [==============================] - 1s 563us/step - loss: 0.6441\n",
      "Epoch 7/20\n",
      "1769/1769 [==============================] - 1s 583us/step - loss: 0.6403\n",
      "Epoch 8/20\n",
      "1769/1769 [==============================] - 1s 587us/step - loss: 0.6366\n",
      "Epoch 9/20\n",
      "1769/1769 [==============================] - 1s 604us/step - loss: 0.6325\n",
      "Epoch 10/20\n",
      "1769/1769 [==============================] - 1s 726us/step - loss: 0.6267\n",
      "Epoch 11/20\n",
      "1769/1769 [==============================] - 1s 811us/step - loss: 0.6192 0s - los\n",
      "Epoch 12/20\n",
      "1769/1769 [==============================] - 2s 914us/step - loss: 0.6106\n",
      "Epoch 13/20\n",
      "1769/1769 [==============================] - 2s 918us/step - loss: 0.6017\n",
      "Epoch 14/20\n",
      "1769/1769 [==============================] - 2s 930us/step - loss: 0.5880\n",
      "Epoch 15/20\n",
      "1769/1769 [==============================] - 2s 894us/step - loss: 0.5752\n",
      "Epoch 16/20\n",
      "1769/1769 [==============================] - 2s 912us/step - loss: 0.5571\n",
      "Epoch 17/20\n",
      "1769/1769 [==============================] - 2s 856us/step - loss: 0.5405\n",
      "Epoch 18/20\n",
      "1769/1769 [==============================] - 2s 897us/step - loss: 0.5180\n",
      "Epoch 19/20\n",
      "1769/1769 [==============================] - 2s 857us/step - loss: 0.4980\n",
      "Epoch 20/20\n",
      "1769/1769 [==============================] - 1s 846us/step - loss: 0.4792\n",
      "Sequence Length of 9\n",
      "Layer size 200\n",
      "Epoch 1/20\n",
      "1708/1708 [==============================] - 6s 3ms/step - loss: 0.6775\n",
      "Epoch 2/20\n",
      "1708/1708 [==============================] - 2s 994us/step - loss: 0.6609\n",
      "Epoch 3/20\n",
      "1708/1708 [==============================] - 1s 834us/step - loss: 0.6556\n",
      "Epoch 4/20\n",
      "1708/1708 [==============================] - 1s 852us/step - loss: 0.6512\n",
      "Epoch 5/20\n",
      "1708/1708 [==============================] - 1s 841us/step - loss: 0.6480\n",
      "Epoch 6/20\n",
      "1708/1708 [==============================] - 1s 876us/step - loss: 0.6437\n",
      "Epoch 7/20\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1708/1708 [==============================] - 1s 659us/step - loss: 0.6400\n",
      "Epoch 8/20\n",
      "1708/1708 [==============================] - 1s 635us/step - loss: 0.6359\n",
      "Epoch 9/20\n",
      "1708/1708 [==============================] - 1s 635us/step - loss: 0.6306\n",
      "Epoch 10/20\n",
      "1708/1708 [==============================] - 1s 705us/step - loss: 0.6250\n",
      "Epoch 11/20\n",
      "1708/1708 [==============================] - 1s 622us/step - loss: 0.6171\n",
      "Epoch 12/20\n",
      "1708/1708 [==============================] - 1s 621us/step - loss: 0.6066 0s\n",
      "Epoch 13/20\n",
      "1708/1708 [==============================] - 1s 657us/step - loss: 0.5944\n",
      "Epoch 14/20\n",
      "1708/1708 [==============================] - 1s 730us/step - loss: 0.5832\n",
      "Epoch 15/20\n",
      "1708/1708 [==============================] - 1s 621us/step - loss: 0.5671 0s\n",
      "Epoch 16/20\n",
      "1708/1708 [==============================] - 1s 613us/step - loss: 0.5507\n",
      "Epoch 17/20\n",
      "1708/1708 [==============================] - 1s 617us/step - loss: 0.5298\n",
      "Epoch 18/20\n",
      "1708/1708 [==============================] - 1s 620us/step - loss: 0.5126\n",
      "Epoch 19/20\n",
      "1708/1708 [==============================] - 1s 687us/step - loss: 0.4908 0s - loss:\n",
      "Epoch 20/20\n",
      "1708/1708 [==============================] - 1s 613us/step - loss: 0.4737\n"
     ]
    }
   ],
   "source": [
    "for i in range(max_sequence_length):\n",
    "    print(\"Sequence Length of \" + str(i))\n",
    "    for size in layer_sizes:\n",
    "        print(\"Layer size \" + str(size))\n",
    "        key = str(size)+'_neuron_layers'       \n",
    "        # define the LSTM model\n",
    "        model = Sequential()\n",
    "        model.add(LSTM(size, input_shape=(i+1, nteams)))\n",
    "        model.add(Dropout(0.2))\n",
    "        #model.add(Dense(nteams, activation='softmax'))\n",
    "        model.add(Dense(nteams, activation='tanh'))\n",
    "        #model.compile(loss='categorical_crossentropy', optimizer='adam')\n",
    "        model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "        histories[key].append(model.fit(X_trainings[i], np.array(Y_trainings[i]), epochs=epochs))\n",
    "        rnns[key].append(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "del histories['200_neuron_layers'][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'200_neuron_layers': [<keras.callbacks.History at 0x1a5866ac88>,\n",
       "  <keras.callbacks.History at 0x1a4b6eeef0>,\n",
       "  <keras.callbacks.History at 0x1a5a4d5c88>,\n",
       "  <keras.callbacks.History at 0x1a58590f98>,\n",
       "  <keras.callbacks.History at 0x1a5c7a9ba8>,\n",
       "  <keras.callbacks.History at 0x1a5e6decc0>,\n",
       "  <keras.callbacks.History at 0x1a4962db70>,\n",
       "  <keras.callbacks.History at 0x1a6027ac18>,\n",
       "  <keras.callbacks.History at 0x1a5bb1bf98>,\n",
       "  <keras.callbacks.History at 0x1a63ab2c88>]}"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "histories"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'loss': [0.68784040405415414,\n",
       "  0.66445436784531298,\n",
       "  0.66005649181610193,\n",
       "  0.65749570662646784,\n",
       "  0.65578104974846418,\n",
       "  0.6546577947510519,\n",
       "  0.65359442567846449,\n",
       "  0.65292801387446298,\n",
       "  0.65249176399339126,\n",
       "  0.65128319658005618,\n",
       "  0.65035874393591586,\n",
       "  0.64926546856952438,\n",
       "  0.64824505532585419,\n",
       "  0.64770746880735397,\n",
       "  0.64588524228870525,\n",
       "  0.64485777438505598,\n",
       "  0.64390761972688493,\n",
       "  0.64238836219155482,\n",
       "  0.64096725581227865,\n",
       "  0.6395478477032247]}"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "histories['200_neuron_layers'][0].history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAD8CAYAAAB3u9PLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a663b5748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(histories['200_neuron_layers'][7].history['loss'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def score(key):\n",
    "    training_accs = []\n",
    "    training_errs = []\n",
    "\n",
    "    print(key)\n",
    "    with open('rnn_training_scores_' + key + '.csv', 'w') as f:\n",
    "        writer = csv.writer(f)\n",
    "        writer.writerow(['Games before prediction', 'Accuracy', 'Euclidean Error'])\n",
    "        print(\"training\")\n",
    "        for i in range(max_sequence_length):\n",
    "            print(i)\n",
    "            outputs = rnns[key][i].predict(X_trainings[i])\n",
    "            acc = 1 - disc_error(outputs, np.array(Y_trainings[i]))\n",
    "            eerr = eucl_error(outputs, np.array(Y_trainings[i]))\n",
    "\n",
    "            writer.writerow([i, acc, eerr])\n",
    "            training_accs.append(acc)\n",
    "            training_errs.append(eerr)\n",
    "            \n",
    "    testing_accs = []\n",
    "    testing_errs = []\n",
    "\n",
    "    with open('rnn_testing_scores_' + key + '.csv', 'w') as f:\n",
    "        writer = csv.writer(f)\n",
    "        writer.writerow(['Games before prediction', 'Accuracy', 'Euclidean Error'])\n",
    "        print(\"testing\")\n",
    "        for i in range(max_sequence_length):\n",
    "            print(i)\n",
    "            outputs = rnns[key][i].predict(X_testings[i])\n",
    "            acc = 1 - disc_error(outputs, np.array(Y_testings[i]))\n",
    "            eerr = eucl_error(outputs, np.array(Y_testings[i]))\n",
    "\n",
    "            writer.writerow([i, acc, eerr])\n",
    "            testing_accs.append(acc)\n",
    "            testing_errs.append(eerr)\n",
    "    \n",
    "    training_accs_dict[key] = training_accs\n",
    "    training_errs_dict[key] = training_errs\n",
    "    testing_accs_dict[key] = testing_accs\n",
    "    testing_errs_dict[key] = testing_errs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "training_accs_dict = {}\n",
    "training_errs_dict = {}\n",
    "testing_accs_dict = {}\n",
    "testing_errs_dict = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{}"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_errs_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "200\n",
      "200_neuron_layers\n",
      "training\n",
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n",
      "9\n",
      "testing\n",
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n",
      "9\n"
     ]
    }
   ],
   "source": [
    "for size in layer_sizes:\n",
    "    print(size)\n",
    "    key = str(size)+'_neuron_layers'\n",
    "    score(key)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot(y, ylabel, title, imname):\n",
    "    x = range(1,max_sequence_length+1)\n",
    "    yfit = np.poly1d(np.polyfit(x, y, 1))(np.unique(x))\n",
    "    r2 = np.round(r2_score(y, yfit),decimals=2)\n",
    "    mean = np.round(np.mean(y), decimals=2)\n",
    "\n",
    "    plt.scatter(x, y, s = 5)\n",
    "    plt.plot(x, yfit)\n",
    "    plt.axis([0,max_sequence_length+1,0,1])\n",
    "    plt.text(1, 0.95, 'R^2: ' + str(r2))\n",
    "    plt.text(1, 0.85, 'Mean: ' + str(mean))\n",
    "    plt.xlabel('Games before prediction')\n",
    "    plt.ylabel(ylabel)\n",
    "    plt.title(title)\n",
    "    plt.savefig('rnnplots/' + imname + '.png')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a664872e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMDOzjJOCmZll/j8zDPUgRfEroQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a5b9ca2e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a3eec0588>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3XmYFeWZ/vHvza7NpgJGAReCiIgI\n2m4xLtEYkYmoiWYwRo1GDcm4xIxmkvmpUSeZIYljjIkmMcS4JANxDU4konGJ2VBxFxw3RG3U2LIJ\nLdDd9PP7o+oUh+7T3acbTh+6uT/XdS5qeavqqdOHet56q+otRQRmZmYA3codgJmZbT6cFMzMLOOk\nYGZmGScFMzPLOCmYmVnGScHMzDJOClYykkZJWl7uOMyseE4KZSZpkaTVklZJelfSTZL65s2/SVJI\n2j9v2khJkTf+iKQ1kobnTfukpEUFtrdTuq3cJyTV5I0fshH78q6kj+fGI+LliBjY3vUVuc1p6T7s\nXcrtlJOkayW9JmmlpAWSTm40fz9Jz0j6UNLjksbmzesm6YeSlkl6X9J3WtjOREmvNjNvF0m/S9ex\nQtJzkj6f/s5yv52a9G+R//saImluOn33Ruu8L51+YDPbnCnpkrZ9W7axnBQ2D8dGRF9gPDAB+Faj\n+UuBZv8zp2qAS1vbUES8GRF9c5908t550/7c1uDLRVI34Ask389pHb3tdPsd4QPgGGAAcA7wM0n7\npnFsBcwCbgC2AW4H7pbUI132POAoYAywD/DPkr7YjhhmAC8Bw4FBwBnA+xHxx7zf0r7AuvzfV0S8\nly7/Mnl/I0kfAfYCVrQjlg6R9x1uUZwUNiMR8S4whyQ55LsZGCfpsBYWvxY4WdLIjY1D0laSrpH0\nVlr7/7Gk3um8j6Q1vOWSlkh6KJ1+OzAEuD+tIZ4vabSk+rz1zpX07fTfDyTNlrRN3vyzJL0pqVrS\nNxqfeRTwSZID5b8Cp0jq3mg/virp/9Ia9vOS9kqn7yJpVlrrfV/Sf6fTp0manrd8ofivlPQY8CGw\no6Qv523jVUlnNorhpLRW/YGkVyQdKelUSX9tVO7/SZpZaCcj4pL0rKshIv4CPAbkatdHAWsi4vqI\nWAv8N9APyH1vpwPfj4h3IuJN4Brgiy18p01IElAJ/CoiVkdEXUQ8GRH3t2E1vyb5Gykd/wJwG1Df\n/CItxvRTSVXp9/p47mxDyZlwjaT+eWUPlvR2Lomnf7OXJC2VdK+koen0PumZy1ckvQa80J7YOjsn\nhc2IpGEkNcLGp/AfAv8JfLeFxRcDvwAu3wSh/BAYRlKT2x0YBXwznfdvJDXGQcAOue1FxEnAe8Cn\n0hritc2s+/PAKemyA4ELACSNB64GPpdue1i6jZacDtwN/BbYGjg6N0PSqWmsJwP9gROBZZJ6An8A\nXgR2Iqn53tnKdvJ9gaTG2w94F3iH5G/WH5gKXCdpzzSGQ0hq8Bek+3ok8BZwF7CXpBF56z0FuLW1\njStpWtwHmJ9O2hN4Njc/IhpIDmZ7ppPG5M9Ph/ekDSLpC+cx4OeSPpf+TtvqdeBN4PB0/FTglnas\nJ+fvJL/P7UjOlG6X1DNNfI8Bn80r+wXgNxHRIGkK8DXgWGB74GmShJXv0yRnPRM2Ir5Oy0lh8/A7\nSStJDhjvAd8uUObnwE6SjmlhPf8FHJs7KLVHesp8JnBBRCyPiBXANGBKWqQO2BHYKSJqI+LRNm7i\nFxHxWkTUAHew/qzoc8CdETE3rfFeQgu/z7QmeALwPxGxmiQ5nJ5X5CzgPyPi6Ui8FBFVJDXo/sC/\nR8SHac33b22If3q6rrqIqI+IeyLi9XQbfwT+xPpa+lnAzyLi4bSW/2Za468hSUSnpPtSSZIA57S0\n4bSWPR34S0Q8kk7uS9MmmBVAvzQB9m40fwVJQmur44HHgSuANyXNk9TWg+YtwGlpBaBbRDzdjjgA\niIhbImJZRNSRVJi2A3JJ9maSRICkXiS/rVzC/TLwnfTvUJfuz8clbZ+3+u+mv/3V7Y2vM3NS2Dwc\nHxH9SGpRoylQQ04PlP+RftR4flqmGvgJcOVGxLIj0BOYnzYRLQd+R9I0BMnZytvAw2lzydfbuP53\n84Y/JDmo5bb7Vm5GRHxAy+3NJwErgT+m478BjpOUu7A9HHitwHLDgdfTGnV7vJU/Imly2nyxNP2u\njmD936+5GCA5cJ2SDn8BmBERrTWlXAvsnJbPWUWS5PL1B1amB721jeb3J/ne2iQi3o+IiyNiD+Aj\nJNcI7mrjam4HJpGcUW3MWQKSvpU2Aa0AlgF9WP+93wnslzYLTQKqIuK5dN7OJNdkcr/tapImrPyz\nnw3+xlsaJ4XNSET8CbgJuKqZIr8iaUM/oYXV/AD4BMnpb3u8Q/Kf5KMRMTD9DIiI7dIYV0TEBRGx\nM8kp+iWSDs7tQju3mdtu9h8zPRMY0EL500maZKokvUtSE+xNUiuE5D/2Rwss9xawiwpfJK4haYbK\n+UiBMvl3fVWQHOj+AxiS3mn1EOuTdnMxQHJG0SdtC59CK01HkqaRnIEcExGr8mbNB/bOK9cNGMv6\n5qUF+fPT4flshPTi8dUk32NFG5ZbATwMfIkkibeLpKNILqCfQPIb2BZYTfq9p9/P3SRNlaey4Xf7\nFvDFvN/2wIjYKiKezA+1vbF1BU4Km59rgKPSU+wNpDXJy0nayguKiOUkFxu/0Z6Np7XLG4EfSRqk\nxPD0P2KuZrxr2pSxAliXfgD+wfpT+La6Dfisktsre5Gc7RSszadt8R8HPkXS/DSe5GB3DeubkKYD\n35S0d7oPo9K28L+Q1JT/Q9LWSi6qfyxd5hngE5KGKrkA3uz3nNqK5KzqPaBB0mTWt5nnYviypEOV\n3K00XNIoyNrpbyW55rA0IuY1txFJVwDHkVyvafzcxwPAVpKmKrkZ4EKS5PaXdP4twMVKbhAYTtKe\nflML+6T0gmv2SSdeJWmMpO6SBpDU9l9Im8La4iLgsIh4u8jyPRrF05Ok+auOpJaf+630abTcLSTN\ndxPZMAH9jKQis3u6X9tI+iyWcVLYzKRNQLfQ/O2lM0hq1S35EesP1O3xNZImonkkB/77gNxdTXsA\nj5AcWB8FroqIuem87wLfTU/Nz23LBtP25YtJaniLSfZxBUnzR2OnAX+PiD9FxLu5D8l+HyBpt4i4\nlaQ2ewfJLZ13AAPTpDeJJIlUkVz8/Ey63nuB35PUrueSNJu1FPP7JAe5/wWWkLS7z86b/2eSg+f1\n6b48yIbNFDeTXCxt9iwhPdBfRpJsX9f6+/+/nm5jNUnCmAosJznrOD6vKeradLsvkiS92yPiphZ2\nawRJrTv7pMm0P3BPuh+vAoNZ/70VLSKq2ngN59uN4vkDyff9KEnT3ELgfZIEke9hkqT9l4jI/r9E\nxAySJta7JH1A8p0c1db96MoUfsmObYbSmvpSYMf8/9RdiaR+JGdXo9O7ZmwTkvQ34PqIaHx3kbXA\nZwq22UibprZKb7u8GnisqyaE1HnAI04Im156nWsUbbvd2ChhUpB0o6T3JBV8ACRt5702vYPlOUn7\nlCoW6zROIrk7qQoYyvq7c7qc9OL4WbTz2o81T8lDgL8Hzt9SbyvdGCVrPpJ0KMntcrdExNgC8yeR\n1JQmAQcAP4qIA0oSjJmZFaVkZwrpQ01LWyhyHEnCiPRC5UBJO5QqHjMza105O3wayoYPieSaDJq0\nIUs6h6QjMCoqKvYdPXp0hwRoZtZVPPnkk+9HxODWypUzKRR6KrdgW1ZE3EByPzeVlZUxb16zt3Sb\nmVkBkt4oplw57z6qIukGIGcYyb3xZmZWJuVMCveQdI6l9FH/FV389kMzs81eKW9JnUHSve3uSvo9\n/1L6KP7UtMhskqcRXyXp8vmrpYplU+revTvjx49n7NixHHvssSxfvmGvA6tWraKyspIRI0bw9tsb\nnviccsop7L777owdO5YzzzyTurq6Vrd33333sfvuuzNy5EimTZtWsMyFF17I+PHjGT9+PKNGjWLg\nwPUvO3vzzTf51Kc+xR577MGYMWNYtGhR23fazLYcEdGpPvvuu2+UU0VFRTZ82mmnxXe+851svK6u\nLo455pi45ppr4o477ojKyspYsWJFNv/ee++NhoaGaGhoiClTpsT111/f4rbq6+tjxIgR8dprr8Xa\ntWtj3LhxMX/+/BaXufbaa+OMM87Ixg877LC4//77IyJi5cqVUVNT06b9NbOuAZgXRRxjt8jXzW0q\nBx10EM8991w2/uUvf5ljjjmG8847D0jOKqZMmcKsWbPo2bMnkyZNysruv//+VFVVtbj+xx9/nJEj\nRzJiRNLHXG5dY8aMaXaZGTNmcMUVVwCwYMEC6uvrOeqopGuXvn37NrucmRm4m4t2W7duHQ8++CCT\nJ0/Opv3yl7/MEgLA8ccfz+zZs+nZs+cGy9bV1XHrrbcyceJEAObNm8dZZ53VZBuLFy9m+PD11+KH\nDRvG4sWLm43pjTfe4PXXX+eII44A4OWXX2bgwIF85jOfYcKECVx88cWsW7cx/eSZWVfnpNBGq1ev\nZvz48Wy33XYsXbo0q4W3xVe/+lUOPfRQDjnkEAAqKyuZPn16k3JR4GlzqeD7dQCYOXMmJ554It27\nJ68qrq+v589//jNXXXUVTzzxBAsXLuSmm25qc7xmtuVwUmijrbbaimeeeYY33niD2tparrvuujYt\nf8UVV1BdXc3VV1/datlhw4bx1lvrn++rqqpixx13bLb8zJkzOfnkkzdYfsKECYwYMYIePXpw/PHH\n89RTT7UpXjPbsjgptNOAAQO49tprueqqq4q6iwhg+vTpzJkzhxkzZtCtW+tf/X777ccrr7zC66+/\nTm1tLTNnztyguSrfSy+9xLJlyzjooIM2WH7ZsmVUVyddzT/00EMtXo8wM3NS2AgTJkxg7733ZubM\nmUWVnzp1Kv/4xz846KCDGD9+PFdembxKublrCj169OAnP/kJRx99NHvssQef+9zn2HPPPQG47LLL\nuOeee7KyM2bMYMqUKRs0L3Xv3p2rrrqKI488kr322ouI4Oyzz96YXTazLq7TvWTH3VyYmbWdpCcj\norK1cj5TMDOzjJOCmZllnBTMzCzjpGBmZhknBTMzyzgpmJlZxknBzMwyTgpmZpZxUjAzs4yTgpmZ\nZZwUzMws46RgZmYZJwUzM8s4KbSRJE499dRsvL6+nsGDB/PpT3+6w2KICM4//3xGjhzJuHHjmn1x\nzsSJE9l7773Zc889mTp1avYqzmeffZaDDjqIvfbai2OPPZYPPvigw2I3s82bk0IbVVRU8MILL7B6\n9WoAHnjgAYYOHdqhMfzhD3/glVde4ZVXXuGGG27gK1/5SsFyt912G88++ywvvPAC1dXV3H777QCc\nddZZTJs2jeeff54TTjiBH/zgBx0ZvpltxpwU2uGYY47h3nvvBZKX2+S/ArOmpoYzzzyT/fbbjwkT\nJjBr1iwAFi1axCGHHMI+++zDPvvsw9/+9jcAHnnkEQ4//HBOPPFERo8ezSmnnFLw3cz5Zs2axWmn\nnYYkDjzwQJYvX84777zTpFz//v2B5GymtrY2ewHPSy+9xKGHHgrAUUcdxZ133rmR34iZdRVOCu0w\nZcoUZs6cyZo1a3juuec44IADsnnf/e53OeKII3jiiSd4+OGHufjii6mpqWHIkCE88MADPPXUU/z2\nt7/l/PPPz5Z5+umnueaaa1iwYAELFy7kr3/9K9D07Wo5ixcvZvjw4dn4sGHDWLx4ccFYjz76aIYM\nGUK/fv048cQTARg7dmy23ttvv32D90Cb2ZbNSaEdxo0bx6JFi5gxYwaTJk3aYN7999/PtGnTGD9+\nPIcffjhr1qzhzTffpK6ujrPPPpu99tqLk046iQULFmTL7L///gwbNoxu3boxfvx4Fi1aBMCVV15Z\n8J3Mhc4k8l/DmW/OnDm88847rF27loceegiAG2+8keuuu459992XlStX0qtXr/Z+FWbWxfQodwCd\n1eTJk7nooot45JFHWLJkSTY9IrjzzjvZfffdNyh/+eWXs/322/Pss8/S0NBAnz59snm9e/fOhrt3\n7059fX2L2x42bNgGtfuqqip23HHHZsv36dOHyZMnM2vWLI466ihGjx7N/fffD8DLL7+cNYWZmflM\noZ3OPPNMLrvsMvbaa68Nph999NH8+Mc/zmrzTz/9NAArVqxghx12oFu3btx6663ZnUDtMXnyZG65\n5RYigrlz5zJgwAB22GGHDcqsWrUqu85QX1/P7NmzGT16NADvvfceAA0NDXznO99h6tSp7Y7FzLoW\nJ4V2GjZsGBdccEGT6Zdeeil1dXWMGzeOsWPHcumllwLw1a9+lZtvvpkDDzyQl19+mYqKila30dw1\nhUmTJjFixAhGjhzJ2WefzfXXX5/NGz9+PJBc8J48eTLjxo1j7733ZsiQIdnBf8aMGYwaNYrRo0ez\n4447csYZZ7TrOzCzrket3emyuamsrIx58+aVOwwzs05F0pMRUdlaOZ8pmJlZxknBzMwyTgpmZpZx\nUjAzs4yTgpmZZZwUzMwsU9KkIGmipJckvSrpmwXm7yTpYUlPS3pO0qRC6zEzs45RsqQgqTtwHXAM\nMAY4WdKYRsUuAW6LiAnAFOB6zMysbEp5prA/8GpELIyIWmAmcFyjMgH0T4cHAG+XMB4zM2tFKZPC\nUCC/T+aqdFq+y4EvSKoCZgPnFVqRpHMkzZM0r7q6uhSxmpkZpU0KhfpybtynxsnATRExDJgE3Cqp\nSUwRcUNEVEZE5eDBg0sQqpmZQWmTQhUwPG98GE2bh74E3AYQEX8H+gCDShiTmZm1oJRJ4QlgN0m7\nSupFciG5cZefbwJHAkjagyQpuH3IzKxMSpYUIqIeOBeYA7xIcpfRfElXSsq9TuxfgbMlPQvMAL4Y\nna3bVjOzLqSkb16LiNkkF5Dzp12WN7wAOLiUMZiZWfH8RLOZmWWcFMzMLOOkYGZmGScFMzPLOCmY\nmVnGScHMzDJOCmZmlnFSMDOzjJOCmZllnBTMzCzjpGBmZhknBTMzyzgpmJlZxknBzMwyTgpmZpZx\nUjAzs4yTgpmZZZwUzMws46RgZmYZJwUzM8s4KZiZWcZJwczMMk4KZmaWcVIwM7OMk4KZmWWcFMzM\nLOOkYGZmGScFMzPLOCmYmVnGScHMzDJOCmZmlnFSMDOzjJOCmZllnBTMzCxT0qQgaaKklyS9Kumb\nzZT5nKQFkuZL+p9SxmNmZi3rUaoVS+oOXAccBVQBT0i6JyIW5JXZDfgWcHBELJM0pFTxmJlZ60p5\nprA/8GpELIyIWmAmcFyjMmcD10XEMoCIeK+E8ZiZWStKmRSGAm/ljVel0/KNAkZJ+qukuZImFlqR\npHMkzZM0r7q6ukThmplZKZOCCkyLRuM9gN2Aw4GTgemSBjZZKOKGiKiMiMrBgwdv8kDNzCzRalKQ\ndK6kbdqx7ipgeN74MODtAmVmRURdRLwOvESSJMzMrAyKOVP4CMlF4tvSu4kKnQEU8gSwm6RdJfUC\npgD3NCrzO+ATAJIGkTQnLSxy/WZmtom1mhQi4hKS2vsvgS8Cr0j6T0kfbWW5euBcYA7wInBbRMyX\ndKWkyWmxOcASSQuAh4GLI2JJu/fGzMw2SlG3pEZESHoXeBeoB7YB7pD0QER8o4XlZgOzG027LH+9\nwNfTj5mZlVmrSUHS+cDpwPvAdJLafJ2kbsArQLNJwczMOpdizhQGAZ+JiDfyJ0ZEg6RPlyYsMzMr\nh2IuNM8GluZGJPWTdABARLxYqsDMzKzjFZMUfgqsyhuvSaeZmVkXU0xSUHpBGEiajShhn0lmZlY+\nxSSFhZLOl9Qz/VyAnyUwM+uSikkKU4GPAYtJnkA+ADinlEGZmVl5tNoMlPZcOqUDYjEzszIr5jmF\nPsCXgD2BPrnpEXFmCeMyM7MyKKb56FaS/o+OBv5E0rHdylIGZWZm5VFMUhgZEZcCNRFxM/BPwF6l\nDcvMzMqhmKRQl/67XNJYYACwS8kiMjOzsinmeYMb0vcpXELS9XVf4NKSRmVmZmXRYlJIO737IH2H\n8qPAiA6JyszMyqLF5qP06eVzOygWMzMrs2KuKTwg6SJJwyVtm/uUPDIzM+twxVxTyD2P8C950wI3\nJZmZdTnFPNG8a0cEYmZm5VfME82nFZoeEbds+nDMzKycimk+2i9vuA9wJPAU4KRgZtbFFNN8dF7+\nuKQBJF1fmJlZF1PM3UeNfQjstqkDMTOz8ivmmsL/ktxtBEkSGQPcVsqgzMysPIq5pnBV3nA98EZE\nVJUoHjMzK6NiksKbwDsRsQZA0laSdomIRSWNzMzMOlwx1xRuBxryxtel08zMrIsp5kyhR0TU5kYi\nolZSrxLGZGbWaUUEyz+so3rVWt5fuZbqVWupXrn+842Jo/nIgD6tr6hMikkK1ZImR8Q9AJKOA94v\nbVhmZpuXmrX1yYG90UH+/dz4qvXjdeuiyfK9enRjcN/eLPuwttMnhanAbyT9JB2vAgo+5Wxm1pms\nrV/H+6tqk4N5gQN+/oH+w9p1TZbv3k1sV9GLwf16M7hfb3bfvh+D+vVmcN/e2bTcp1/vHkgqw162\nTTEPr70GHCipL6CI8PuZzWyjNTQES2pqGdS31yY9WK5rCJbW1BZVq1+xuq7gOrbZuieD0gP7hJ0G\nNjnI5+Zts3Uvunfb/A/0bVHMcwr/CXw/Ipan49sA/xoRl5Q6ODPrmhoagpN/MZcn31jGvjtvw4yz\nD6RbCwfXiOCD1fVUr1rDe9kBvrZJjb565VqW1qyloWnrDRW9umcH9J223ZqPfXS7gjX67Sp606tH\ne57rbX2fS5EEN7Vimo+OiYh/z41ExDJJk0hez2lm1mZLamp58o1l1DcE8xYt5aGX3qN+XUPTg/yq\n2qRZZ+Vaatc1NFlPz+7KDuxDB/Zh/PABGxzoczX6QX17U9G7R5aMHn25mn133obLz96zxWS0qbQ1\nCZZTMUmhu6TeEbEWkucUgN6lDctsy9JZapHFqK1vYElNo+aaAjX6hkiq8+sCzrp5Xra8BNtV9MoO\n6B8dXJEc5PMO9kPSA/2ArXq26fvKT0ZPvrGMJTW1DO5X+sNZubbbHsUkhV8DD0r6VTp+BnBz6UIy\n27J0hlpkQ0Ow7MPaVtvoq1euZdmHhdvp+/fpkdXaxw4dwOGjhlDRpwe7bLf1Bs03227dix7dN33z\nDcCgvr3Yd+dtsu96UN+Oubu+XNttj2IuNH9f0nPAJwEB9wE7lzowsy1FuWqREcGq3G2WzV2UXbW+\n/X5dgYb6Pj27ZbX4XQdVsP+u2zK4b5/04N9rg2acPj27l3yfWiOJGWcf2OFnZeXabnsUc6YA8C7J\nU82fA14H7ixmIUkTgR8B3YHpETGtmXInkjwlvV9EzCtUxrYcXakppRibuha5pm5ddpBv7TbLNXVN\n2+l7dFPWdDOkXx/G7NA/r/mmzwa1+ope3Tvd36hbN5Wl6aZc222rZpOCpFHAFOBkYAnwW5JbUj9R\nzIoldQeuA44iebbhCUn3RMSCRuX6AecDj7VrD6xLKWdTSrmSUTG1yPp1DSytqU3uvGmlCWflmvqC\n28lvp99ll6SdflBFL3r37M6IwRUMSQ/4A7fqudk1X1nHaelM4f+APwPHRsSrAJIubMO69wdejYiF\n6bIzgeOABY3K/QfwfeCiNqzbuqhyNaWUIxnld4fQ2kXZpR/WEgVus+zXO22n79ebPXboz6G79W5y\nUXZwv95sW9GLno3a6TvDtQzreC0lhc+SnCk8LOk+YCbJNYViDQXeyhuvAg7ILyBpAjA8In4vqdmk\nIOkc4ByAnXbaqQ0hWGdTrgtymyoZRQQ1tesKN9vkmnSK6A5hSHowH77t1uyz8zZN76fvm7TTb9Wr\n/e30nemOGOs4zSaFiLgbuFtSBXA8cCGwvaSfAndHxP2trLtQAsn+B0jqBvwQ+GJrQUbEDcANAJWV\nlQXqS9ZVlOuCXGvJKL87hCbNNo0SwOq6pt0hdBNZ002uO4TG99EP6d+x3SF0pjtirOMoCp2TNldY\n2hY4CfjniDiilbIHAZdHxNHp+LcAIuK/0vEBwGvAqnSRjwBLgcktXWyurKyMefN8Ldo23rqGYEnN\nWt5fmdxq+d4Ha3hjyYd8WFtP9apaqleuyRJBc90hDNy654a1+AIPTm3O3SFsaRf1t2SSnoyIytbK\nFXv3EQARsRT4efppzRPAbpJ2BRaTNEV9Pm9dK4BBeQE/Alzku49sYzTuDqFx7b6Y7hC2TrtDGNy3\nN7sN6dvh3SF0pM5yR4x1nDYlhbaIiHpJ5wJzSG5JvTEi5ku6EpiX64rbNk+bWw1ydW3uNss1eQf4\nDfu+aUt3CHsPG9Bs7b6id8n+W5ht9kr664+I2cDsRtMua6bs4aWMxYrXUXel1K1rYEnuwJ5/sM+v\n4ac1+1Vrm95mmXSHsP4hqSbdIeQd7NvaHYLZlspVImtiY+5KKdQdQnMXZJvrDqFfnx7Z3TdjhyYd\nnA3q16vpbZYl7A7BbEvlpGBNNL4rZbuKnnywpi5rnin0hOz7RXSHMKRfHwb17dWkO4T1TTe9Npvu\nEMy2VE4KW7D87hAa1+gHbNWTPXboz+Llq9njsjmsrW+9O4Q9dxhQ8M6bztodgtmWyEmhi9lU3SHk\nDuwjh/Qt+ITs4LTbYj8Ba9a1OCl0AhHBitV1zfZmmV/TX1Kz6btDMLMth5NCGdXkui0uojfL5rpD\nyB3UC3WHMKjv+peRbEx3CGa25XBS2MRq6xuavdumcfPNh7XNd4eQa5PfbUhfKnrnXkTSZ4MDfv8+\nHdMdgpltOZwUirCuIVhaU9vsm6byx4vpDmH88IHpbZZNm2/yu0NwL5Zm1tGcFIrw2MIlfH5609c9\nVPTqnh3YC3WHkKvtb9e3F717tL35xr1YmllHc1Iowm7b9+PK4/bs8O4Q3IulmXW0NvWSujnY0npJ\n3dz6IDKzzqkkvaRax3MvlmbWkXxDupmZZZwUzMws46RgZmYZJ4UiNTQE1SvX0tkuzJuZtYUvNBfB\nD5GZ2ZbCZwpFKPQQmZlZV+SkUITcQ2Q9uskPkZlZl+bmoyJIYsbZB/ohMjPr8pwUiuSHyMxsS+Dm\nIzMzyzgpmJlZxknBzMwyTgqx/mLUAAALAElEQVRmZpZxUjAzs4yTgpmZZZwUzMws46RgZmYZJwUz\nM8s4KZiZWcZJwczMMk4KZmaWcVIwM7NMSZOCpImSXpL0qqRvFpj/dUkLJD0n6UFJO5cyHjMza1nJ\nkoKk7sB1wDHAGOBkSWMaFXsaqIyIccAdwPdLFY+ZmbWulGcK+wOvRsTCiKgFZgLH5ReIiIcj4sN0\ndC4wrITxmJlZK0qZFIYCb+WNV6XTmvMl4A+FZkg6R9I8SfOqq6s3YYhmZpavlEmh0Dsro2BB6QtA\nJfCDQvMj4oaIqIyIysGDB2/CEM3MLF8pX8dZBQzPGx8GvN24kKRPAv8POCwi1pYwHjMza0UpzxSe\nAHaTtKukXsAU4J78ApImAD8HJkfEeyWMxczMilCypBAR9cC5wBzgReC2iJgv6UpJk9NiPwD6ArdL\nekbSPc2szszMOkApm4+IiNnA7EbTLssb/mQpt29mZm3jJ5rNzCzjpGBmZhknBTMzyzgpmJlZxknB\nzMwyTgpmZpZxUjAzs4yTgpmZZZwUzMws46RgZmYZJwUzM8s4KZiZWcZJwczMMk4KZmaWcVIwM7OM\nk4KZmWWcFMzMLOOkYGZmGScFMzPLOCmYmVnGScHMzDJOCmZmlnFSMDOzjJOCmZllnBTMzCzjpGBm\nZhknBTMzyzgpmJlZxknBzMwyTgpmZpZxUjAzs4yTgpmZZZwUzMws46RgZmYZJwUzM8uUNClImijp\nJUmvSvpmgfm9Jf02nf+YpF1KGY+ZmbWsZElBUnfgOuAYYAxwsqQxjYp9CVgWESOBHwLfK1U8ZmbW\nulKeKewPvBoRCyOiFpgJHNeozHHAzenwHcCRklTCmMzMrAU9SrjuocBbeeNVwAHNlYmIekkrgO2A\n9/MLSToHOCcdXSvphZJEvPkaRKPvZAvgfd4ybGn7XM793bmYQqVMCoVq/NGOMkTEDcANAJLmRUTl\nxofXeXiftwze566vM+xvKZuPqoDheePDgLebKyOpBzAAWFrCmMzMrAWlTApPALtJ2lVSL2AKcE+j\nMvcAp6fDJwIPRUSTMwUzM+sYJWs+Sq8RnAvMAboDN0bEfElXAvMi4h7gl8Ctkl4lOUOYUsSqbyhV\nzJsx7/OWwfvc9W32+ytXzM3MLMdPNJuZWcZJwczMMp0qKbTWbUZXI2m4pIclvShpvqQLyh1TR5DU\nXdLTkn5f7lg6gqSBku6Q9H/p3/qgcsdUapIuTH/TL0iaIalPuWPa1CTdKOm9/OeqJG0r6QFJr6T/\nblPOGAvpNEmhyG4zupp64F8jYg/gQOBftoB9BrgAeLHcQXSgHwH3RcRoYG+6+L5LGgqcD1RGxFiS\nG1GKucmks7kJmNho2jeBByNiN+DBdHyz0mmSAsV1m9GlRMQ7EfFUOryS5GAxtLxRlZakYcA/AdPL\nHUtHkNQfOJTkTjwiojYilpc3qg7RA9gqfT5pa5o+w9TpRcSjNH3uKr9rn5uB4zs0qCJ0pqRQqNuM\nLn2AzJf2IDsBeKy8kZTcNcA3gIZyB9JBRgDVwK/SJrPpkirKHVQpRcRi4CrgTeAdYEVE3F/eqDrM\n9hHxDiSVPmBImeNpojMlhaK6xOiKJPUF7gS+FhEflDueUpH0aeC9iHiy3LF0oB7APsBPI2ICUMNm\n2KSwKaXt6McBuwI7AhWSvlDeqCynMyWFYrrN6HIk9SRJCL+JiLvKHU+JHQxMlrSIpHnwCEm/Lm9I\nJVcFVEVE7gzwDpIk0ZV9Eng9Iqojog64C/hYmWPqKP+QtANA+u97ZY6nic6UFIrpNqNLSbsR/yXw\nYkRcXe54Si0ivhURwyJiF5K/70MR0aVrkBHxLvCWpN3TSUcCC8oYUkd4EzhQ0tbpb/xIuvjF9Tz5\nXfucDswqYywFlbKX1E2quW4zyhxWqR0MnAo8L+mZdNq/R8TsMsZkm955wG/Sys5C4Iwyx1NSEfGY\npDuAp0jusHuaTtD9Q1tJmgEcDgySVAV8G5gG3CbpSyTJ8aTyRViYu7kwM7NMZ2o+MjOzEnNSMDOz\njJOCmZllnBTMzCzjpGBmZhknBSsZSdtL+h9JCyU9Kenvkk4oc0y75PdaWeQyoyU9k3ZD8dFSxbYp\nSLpJ0onp8PSWOlCUdLikj+WNT5V0WkfEaZuvTvOcgnUu6UNJvwNujojPp9N2BiaXNbD2OR6YFRHf\nLqZwuu+KiE3Sf5OkHhFR39blIuKsVoocDqwC/paW/1nbo7OuxmcKVipHALX5B5qIeCMifgxZjf3P\nkp5KPx9Lpx8u6U+SbpP0sqRpkk6R9Lik53M1dUmDJd0p6Yn0c3A6/bC0Vp+r2fcrEFsPSTdLei59\nj8HW6bL7ptt+UtIcSTtImgR8DThL0sNpua+n7wF4QdLX8vbnRUnXkzyUNVzSp9Kzo6ck3Z72YbUB\nSY9IukbS39L17Z9Ov1zSDZLuB25R8o6JH6T7+pykL6flJOknkhZIupe8DtbSdVemwxPTOJ6V9KCS\nDhanAhem39Uh6TYvSsuPlzQ33dbdSvv9T9f5vfTv8bKkQ9rz47DNWET4488m/5D0l//DFuZvDfRJ\nh3cD5qXDhwPLgR2A3sBi4Ip03gXANenw/wAfT4d3IukKBOB/gYPT4b5Aj0bb3YWkI8VcmRuBi4Ce\nJDXmwen0fyZ5ah7gcuCidHhf4HmgIl3/fJLea3ch6dn1wLTcIOBRoCId/zfgsgLfwyPAL9LhQ4EX\n8rb5JLBVOn4OcEk63BuYR9Kh3GeAB0ie8t8x/e5OzFt3JTCYpIfhXdPp2zberwL7+RxwWDp8Zd73\n/gjw3+nwJOCP5f6t+bNpP24+sg4h6Trg4yRnD/uRHIR/Imk8sA4YlVf8iUi7F5b0GpDrVvl54BPp\n8CeBMUlLDQD907OCvwJXS/oNcFdEVBUI562I+Gs6/GuSBHYfMBZ4IF1nd5JunRv7OHB3RNSk8d0F\nHELSp80bETE3LXcgycug/pqurxfw92a+nhmQ9L8vqb+kgen0eyJidTr8KWBc7noBMIAkmR4KzIiI\ndcDbkh4qsP4DgUcj4vV0O437+N+ApAHAwIj4UzrpZuD2vCK5jhmfJEmG1oU4KVipzAc+mxuJiH+R\nNIikhgtwIfAPkjeNdQPW5C27Nm+4IW+8gfW/2W7AQXkHzZxpaTPKJGCupE9GxP81KtO4b5cg6Zp9\nfkS09irMQl2459Q0KvdARJzcyvqai6fQ+s6LiDkbBJM0b7XWV42KKNMWub/HOnwM6XJ8TcFK5SGg\nj6Sv5E3bOm94APBOJBdjTyWpmbfF/cC5uZH0jANJH42I5yPieyQJaHSBZXfS+vcgnwz8BXgJGJyb\nLqmnpD0LLPsocLySHj4rgBOAPxcoNxc4WNLIdH1bSxpVoBwkTVVI+jjJC2dWFCgzB/iKkq7UkTQq\n3f6jwJT0msMOrD+Tyvd34DBJu6bLbptOXwk0ueaSbn9Z3vWCU4E/NS5nXZOTgpVERATJXTuHSXpd\n0uMkzRD/lha5Hjhd0lySpqOawmtq1vlAZXohdAHJRVOAr6UXbJ8FVgN/KLDsi+m2nwO2JXnBTS1w\nIvC9dNlnKNDHfySvR70JeJzkLXjTI+LpAuWqgS8CM9LtzKVwgoLkAPw34GfAl5opM52kS+2nlNxS\n+3OSWvrdwCskTWs/pcDBO43lHOCudN9+m876X+CE3IXmRoudDvwgjX08yXUF2wK4l1SzMpL0CMnF\n3XmtlTXrCD5TMDOzjM8UzMws4zMFMzPLOCmYmVnGScHMzDJOCmZmlnFSMDOzzP8HtYK3ufq6YNkA\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a46dca940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for size in layer_sizes:\n",
    "    key = str(size)+'_neuron_layers'\n",
    "    title = ' ' + str(size) + ' LSTM Layer'\n",
    "    plot(training_errs_dict[key], 'Euclidian Error', 'RNN Training Euclidean Error' + title, 'rnn_training_error_plot_'+key)\n",
    "    plot(testing_errs_dict[key], 'Euclidian Error', 'RNN Testing Euclidean Error' + title, 'rnn_testing_error_plot_'+key)\n",
    "    plot(training_accs_dict[key], 'Accuracy', 'RNN Training Accuracy' + title, 'rnn_training_accuracy_plot_'+key)\n",
    "    plot(testing_accs_dict[key], 'Accuracy', 'RNN Testing Accuracy' + title, 'rnn_testing_accuracy_plot_'+key)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}