cms_zee_conditional_wgan.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data / MC cluster shape corrections using $Z\\rightarrow ee$ sample."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import GAN.models as models\n",
    "import GAN.cms_datasets as cms\n",
    "import GAN.plotting as plotting\n",
    "import GAN.preprocessing as preprocessing\n",
    "import GAN.base as base"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Configuration parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'AM_OPTS': {'loss': 'wgan_loss',\n",
       "  'opt_kwargs': {'decay': 1e-06, 'lr': 1e-05},\n",
       "  'optimizer': 'RMSprop'},\n",
       " 'BATCH': False,\n",
       " 'BATCH_SIZE': 4096,\n",
       " 'CLEANING': 'cleaning_zee_m_barrel.hd5',\n",
       " 'C_NAMES': ['Phi', 'ScEta', 'Pt', 'rho', 'run'],\n",
       " 'DM_OPTS': {'loss': 'wgan_loss',\n",
       "  'opt_kwargs': {'decay': 1e-06, 'lr': 1e-05},\n",
       "  'optimizer': 'RMSprop'},\n",
       " 'D_OPTS': {'activation': None,\n",
       "  'clip_weights': 0.02,\n",
       "  'do_bn': False,\n",
       "  'kernel_sizes': [1024, 512, 128, 64, 32],\n",
       "  'name': 'D_1024to32x6'},\n",
       " 'EPOCHS': 200,\n",
       " 'FEAT_TRANSFORM': 'gaus',\n",
       " 'FRAC_DATA': 10,\n",
       " 'GAN_TARGETS': 'gan_targets_hinge',\n",
       " 'GLOBAL_MATCH': True,\n",
       " 'G_OPTS': {'do_last_l1reg': 1e-06,\n",
       "  'do_weight_reg': 1e-05,\n",
       "  'kernel_sizes': [64, 64, 128, 128, 256, 256, 512, 512],\n",
       "  'name': 'G_512to64x8'},\n",
       " 'LOAD_DATASETS': ['moriond_v9', 'abs(ScEta) < 1.5'],\n",
       " 'LOSS': 'wgan_loss',\n",
       " 'MCWEIGHT': '',\n",
       " 'MONITOR_DIR': 'log',\n",
       " 'PLOT_EVERY': 10,\n",
       " 'PRETRAIN_G': False,\n",
       " 'REWEIGHT': ['rewei_zee_5d_barrel_m_cleaning_xgb'],\n",
       " 'SAMPLE_FROM_DATA': ['run'],\n",
       " 'SCHEDULE': [0, 0, 0, 0, 0, 1],\n",
       " 'USE_ABS_WEIGHT': True,\n",
       " 'X_NAMES': ['EtaWidth',\n",
       "  'R9',\n",
       "  'SigmaIeIe',\n",
       "  'S4',\n",
       "  'PhiWidth',\n",
       "  'CovarianceIetaIphi',\n",
       "  'CovarianceIphiIphi']}"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import GAN.utils as utils\n",
    "\n",
    "# reload(utils)\n",
    "\n",
    "class Parameters(utils.Parameters):\n",
    "    \n",
    "    # dataset to be loaded\n",
    "    load_datasets=utils.param([\"moriond_v9\",\"abs(ScEta) < 1.5\"])\n",
    "\n",
    "    c_names = utils.param(['Phi','ScEta','Pt','rho','run'])\n",
    "    x_names = utils.param(['EtaWidth','R9','SigmaIeIe','S4','PhiWidth','CovarianceIetaIphi',\n",
    "                           'CovarianceIphiIphi'])\n",
    "    \n",
    "    # generate variables in MC to be distributed like in data\n",
    "    sample_from_data = ['run']\n",
    "    \n",
    "    # MC reweighting\n",
    "    reweight = ['rewei_zee_5d_barrel_m_cleaning_xgb']\n",
    "    mcweight = utils.param('') # if empty ignore original weight\n",
    "    use_abs_weight = utils.param(True) # negative weights are notoriously nasty\n",
    "\n",
    "    # cleaning cuts\n",
    "    cleaning = utils.param('cleaning_zee_m_barrel.hd5')\n",
    "    \n",
    "    # input features transformation\n",
    "    feat_transform = utils.param('minmax')\n",
    "    global_match = utils.param(False) \n",
    "    \n",
    "    # generator and discriminator (aka as 'critic' in wGAN)\n",
    "    g_opts=utils.param(dict(name=\"G_512to64x8\",\n",
    "                            kernel_sizes=[64,64,128,128,256,256,512,512],\n",
    "                            do_weight_reg=1e-5,do_last_l1reg=1e-5\n",
    "                            ))\n",
    "    pretrain_g = utils.param(False)\n",
    "    d_opts=utils.param(dict(name=\"D_1024to32x6\",kernel_sizes=[1024,512,128,64,32],do_bn=False,\n",
    "                        do_dropout=False,hidden_activation=\"relu\",\n",
    "                        clip_weights=2.e-2,activation=None)) # weights clipping and no actication\n",
    "    # optimizers\n",
    "    dm_opts=utils.param(dict(optimizer=\"RMSprop\",opt_kwargs=dict(lr=1e-5,decay=1e-6)))\n",
    "    am_opts=utils.param(dict(optimizer=\"RMSprop\",opt_kwargs=dict(lr=1e-5,decay=1e-6)))\n",
    "        \n",
    "    loss = utils.param(\"wgan_loss\") # use WGAN loss \n",
    "    gan_targets = utils.param('gan_targets_hinge') # hinge targets are 1, -1 instead of 0, 1\n",
    "    schedule = utils.param([0]*5+[1]*1) # number of critic iterations per generator iterations\n",
    "\n",
    "    # schedule training\n",
    "    epochs=utils.param(200)\n",
    "    batch_size=utils.param(4096)\n",
    "    plot_every=utils.param(10)\n",
    "    frac_data=utils.param(10) # fraction of data to use\n",
    "    \n",
    "    monitor_dir = utils.param('log') # folder for logging\n",
    "    \n",
    "    batch = utils.param(False) # are we running in batch?\n",
    "\n",
    "class MyApp(utils.MyApp):\n",
    "    classes = utils.List([Parameters])\n",
    "\n",
    "# Read all parameters above from command line. \n",
    "# Note: names are all converted to be all capital\n",
    "notebook_parameters = Parameters(MyApp()).get_params()\n",
    "\n",
    "# copy parameters to global scope\n",
    "globals().update(notebook_parameters)\n",
    "DM_OPTS.update( {\"loss\":LOSS} )\n",
    "AM_OPTS.update( {\"loss\":LOSS} )\n",
    "\n",
    "\n",
    "plotting.batch = BATCH \n",
    "\n",
    "notebook_parameters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load datasets\n",
    "\n",
    "Apply cleaning and reweight MC to match data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data,mc = cms.load_zee(*LOAD_DATASETS)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['index', 'run', 'rho', 'nvtx', 'mass', 'weight', 'SigMoM', 'Pt',\n",
       "       'ScEta', 'Phi', 'R9', 'S4', 'SigmaIeIe', 'EtaWidth', 'PhiWidth',\n",
       "       'CovarianceIphiIphi', 'SigmaRR', 'ScEnergy', 'CovarianceIetaIphi',\n",
       "       'PhoIso03', 'ChIso03', 'ChIso03worst', 'ScPreshowerEnergy', 'PhoIDMVA',\n",
       "       'SigEOverE', 'run_quantile'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['index', 'run', 'rho', 'nvtx', 'mass', 'weight', 'SigMoM', 'Pt',\n",
       "       'ScEta', 'Phi', 'R9', 'S4', 'SigmaIeIe', 'EtaWidth', 'PhiWidth',\n",
       "       'CovarianceIphiIphi', 'SigmaRR', 'ScEnergy', 'CovarianceIetaIphi',\n",
       "       'PhoIso03', 'ChIso03', 'ChIso03worst', 'ScPreshowerEnergy', 'PhoIDMVA',\n",
       "       'SigEOverE'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mc.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cleaning data and mc\n",
      "cleaning eff (data,mc): 0.91  0.95\n"
     ]
    }
   ],
   "source": [
    "if not CLEANING is None:\n",
    "    print('cleaning data and mc')\n",
    "    thr_up = pd.read_hdf(CLEANING,'thr_up')\n",
    "    thr_down = pd.read_hdf(CLEANING,'thr_down')\n",
    "    nevts_data = data.shape[0]\n",
    "    nevts_mc = mc.shape[0]\n",
    "    data = data[ ((data[thr_down.index] >= thr_down) & (data[thr_up.index] <= thr_up)).all(axis=1) ]\n",
    "    mc = mc[ ((mc[thr_down.index] >= thr_down) & (mc[thr_up.index] <= thr_up)).all(axis=1) ]\n",
    "    print('cleaning eff (data,mc): %1.2f % 1.2f' % (  data.shape[0] / nevts_data, mc.shape[0] / nevts_mc  ))\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate extra variables for MC (eg run number) distributed like data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sampling run\n"
     ]
    }
   ],
   "source": [
    "for col in SAMPLE_FROM_DATA:\n",
    "    print('sampling', col)\n",
    "    mc[col] = data[col].sample(mc.shape[0]).values\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Select target and conditional features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "c_names = C_NAMES\n",
    "x_names = X_NAMES\n",
    "\n",
    "data_c = data[c_names]\n",
    "data_x = data[x_names]\n",
    "\n",
    "mc_c = mc[c_names]\n",
    "mc_x = mc[x_names]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Index(['EtaWidth', 'R9', 'SigmaIeIe', 'S4', 'PhiWidth', 'CovarianceIetaIphi',\n",
       "        'CovarianceIphiIphi'],\n",
       "       dtype='object'),\n",
       " (16024468, 7),\n",
       " Index(['Phi', 'ScEta', 'Pt', 'rho', 'run'], dtype='object'),\n",
       " (16024468, 5))"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_x.columns, data_x.shape, data_c.columns, data_c.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Index(['EtaWidth', 'R9', 'SigmaIeIe', 'S4', 'PhiWidth', 'CovarianceIetaIphi',\n",
       "        'CovarianceIphiIphi'],\n",
       "       dtype='object'),\n",
       " Index(['Phi', 'ScEta', 'Pt', 'rho', 'run'], dtype='object'))"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_x.columns, data_c.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Index(['EtaWidth', 'R9', 'SigmaIeIe', 'S4', 'PhiWidth', 'CovarianceIetaIphi',\n",
       "        'CovarianceIphiIphi'],\n",
       "       dtype='object'),\n",
       " Index(['Phi', 'ScEta', 'Pt', 'rho', 'run'], dtype='object'))"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mc_x.columns, mc_c.columns\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reweight MC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 1.  1.  1.  1.  1.  1.  1.  1.  1.  1.]\n"
     ]
    }
   ],
   "source": [
    "# reload(preprocessing)\n",
    "# reload(utils)\n",
    "\n",
    "# iniliatlize weights to default\n",
    "if MCWEIGHT == '':\n",
    "    mc_w = np.ones(mc_x.shape[0])\n",
    "else:\n",
    "    mc_w = mc[MCWEIGHT].values\n",
    "\n",
    "print(mc_w[:10])\n",
    "    \n",
    "# take care of negative weights\n",
    "if USE_ABS_WEIGHT:\n",
    "    mc_w = np.abs(mc_w)\n",
    "\n",
    "# reweighting\n",
    "if not REWEIGHT is None:\n",
    "    for fil in REWEIGHT:\n",
    "        # apply weights from n-dimnesional histograms\n",
    "        if 'npy' in fil:\n",
    "            info = np.load(fil)\n",
    "            inputs = info[0]\n",
    "            weights = info[1]\n",
    "            bins = info[2:]\n",
    "            # print(bins[1])\n",
    "            print('weighting',inputs)\n",
    "            mc_w *= preprocessing.reweight(mc,inputs,bins,weights,base=None)\n",
    "        else:\n",
    "            # or use a classifier\n",
    "            clf = utils.read_pickle(fil,encoding='latin1')\n",
    "            if hasattr(clf,'nthread'):\n",
    "                clf.nthread = min(4,clf.nthread)\n",
    "                clf.n_jobs = clf.nthread\n",
    "            mc_w *= preprocessing.reweight_multidim(mc[clf.features_],clf)\n",
    "            \n",
    "data_w = np.ones(data_x.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Features transformation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# reload(preprocessing)\n",
    "\n",
    "if GLOBAL_MATCH:\n",
    "    _,data_c,mc_x,mc_c,scaler_x,scaler_c = preprocessing.transform(None,data_c,mc_x,mc_c,FEAT_TRANSFORM)\n",
    "    _,_,data_x,_,scaler_x_data,_ = preprocessing.transform(None,None,data_x,None,FEAT_TRANSFORM)\n",
    "    \n",
    "else:\n",
    "    data_x,data_c,mc_x,mc_c,scaler_x,scaler_c = preprocessing.transform(data_x,data_c,mc_x,mc_c,FEAT_TRANSFORM)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((16024468, 1, 7), (9131160, 1, 7))"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_x.shape,mc_x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1.2261241192018459, 1.3815607851977525)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_x[2].max(),mc_x[2].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 5.19933758  5.19933758  5.19933758  5.19933758  5.19933758]] [[-5.19933758 -5.19933758 -5.19933758 -5.19933758 -5.19933758]]\n",
      "[[ 5.19933758  5.19933758  5.19933758  3.48306001  5.19933758]] [[-5.19933758 -5.19933758 -5.19933758 -5.19933758 -5.19933758]]\n",
      "(9131160,) (9131160, 1, 5)\n"
     ]
    }
   ],
   "source": [
    "print(mc_c.max(axis=0),mc_c.min(axis=0))\n",
    "print(data_c.max(axis=0),data_c.min(axis=0))\n",
    "print(mc_w.shape,mc_c.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAJQCAYAAABM/CoCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2c3f43eeb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAJQCAYAAABM/CoCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTExuYZkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6a4c828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAJQCAYAAABM/CoCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6a89fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAJQCAYAAABM/CoCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0IlgBQDQiWAFANCJYAUA0IlgBQDQyf8HukuW528jzcQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6ac1ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAJQCAYAAABM/CoCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIlgBADTy/wBE1YBx12ybhwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bfcd98080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xu0nWV9L/rvTy5SgYIX2lO5Je6CiIhAE4rXxnpLyxHs\nqZyi7VDG7jH1wlHb4ziN1qFrI45BtwxLdditGZWhezcUrbTbKHgpp0SrlpqgiBKQm1GCVhGNSmnU\nyHP+WNPsZdYMmSQzz1yXz2eMjMz5zPed+a3JCuub3/O8z1uttQAAsG89ZNIFAAAsBkIXAEAHQhcA\nQAdCFwBAB0IXAEAHQhcAQAdCFwBAB0IXAEAHQhcAQAf7T7qAnT3qUY9qS5YsmXQZAAC7dd11132n\ntXbEKMfOudC1ZMmSbNy4cdJlAADsVlV9bdRjTS8CAHQgdAEAdCB0AQB0MOfWdAEAe+4nP/lJtmzZ\nkm3btk26lAXloIMOylFHHZUDDjhgj99D6AKABWTLli059NBDs2TJklTVpMtZEFprueeee7Jly5Ys\nXbp0j9/H9CIALCDbtm3LIx/5SIFrjKoqj3zkI/e6eyh0AcACI3CN3zg+U6ELAKADa7oAYAFbsvrK\nsb7f5ovOHOv77UuXXHJJVq1alYc97GEjn7N+/fpcfPHF+chHPjL2enS6AIB5qbWW+++/f5evX3LJ\nJbnvvvs6VvTAhC4AYKze/OY357GPfWye+tSn5oUvfGEuvvji3H777Vm5cmV+7dd+LU972tNy8803\nJ0nOO++8vOpVr8qTn/zkPOYxj8kHP/jBHe/z1re+NcuXL8/JJ5+cN73pTUmSzZs357GPfWxe/OIX\n56STTsqdd96Zl7/85Vm2bFke//jH7zju7W9/e77xjW/kGc94Rp7xjGckST7xiU/kSU96Uk477bSc\nc845uffee5MkH/vYx3LCCSfktNNOy9///d/vs89F6AIAxmbDhg254oor8sUvfjEf/ehHd9xPedWq\nVXnHO96R6667LhdffHFe8YpX7Djnm9/8Zj796U/nIx/5SFavXp1kOiDdeuut+dznPpfrr78+1113\nXT71qU8lSW699da84hWvyI033phjjz02b3nLW7Jx48bccMMN+eQnP5kbbrghr3rVq/LoRz8611xz\nTa655pp85zvfyYUXXpirr746n//857Ns2bK87W1vy7Zt2/LSl740H/7wh3Pdddfl3/7t3/bZZ2NN\nFwAwNp/5zGdy9tln56CDDspBBx2U5z3vedm2bVs++9nP5pxzztlx3I9+9KMdj5///OfnIQ95SE48\n8cR861vfSjIduj7xiU/k1FNPTZLce++9ufXWW3PMMcfk2GOPzRlnnLHj/A984ANZs2ZNtm/fnm9+\n85vZtGlTTj755J+r69prr82mTZvylKc8JUny4x//OE960pNy8803Z+nSpTnuuOOSJH/wB3+QNWvW\n7JPPRugCAPap+++/P4cffniuv/76oa8/9KEP3fG4tbbj99e97nX5oz/6o587dvPmzTn44IN3PP/q\nV7+aiy++OBs2bMjDH/7wnHfeeUP302qt5dnPfnb+9m//9ufGd1XTvmB6EQAYm6c85Sn58Ic/nG3b\ntuXee+/NRz7ykTzsYQ/L0qVL83d/93dJpgPQF7/4xQd8n+c+97m59NJLd6y7uuuuu/Ltb3971nE/\n+MEPcvDBB+ewww7Lt771rXz0ox/d8dqhhx6aH/7wh0mSM844I5/5zGdy2223JUn+/d//PbfccktO\nOOGEbN68ObfffnuSzApl46TTBQALWO8tHpYvX56zzjorJ598cn75l385T3jCE3LYYYdl7dq1efnL\nX54LL7wwP/nJT3LuuefmiU984i7f5znPeU5uuummPOlJT0qSHHLIIfmbv/mb7Lfffj933BOf+MSc\neuqpOeGEE3L00UfvmD5MpteRrVy5csfarve+97154QtfuGNq88ILL8zxxx+fNWvW5Mwzz8zDHvaw\nPO1pT9sR1MatftbGmyuWLVvWfrboDgB4cG666aY87nGPm2gN9957bw455JDcd999efrTn541a9bk\ntNNOm2hN4zDss62q61pry0Y5X6cLABirVatWZdOmTdm2bVte8pKXLIjANQ5CFwAwVpdddtmkS5iT\nLKQHAOhA6AIA6EDoAgDoQOgCAOjAQnoAWMCm1k+N9/1WPPD7bd26NZdddtnP3VtxX1i/fn0OPPDA\nPPnJT96nf844CV3AvDbsB8rufigA+87WrVvzV3/1VyOHrtZaWmt5yEMe3OTb+vXrc8ghh8yr0GV6\nEQAYm9WrV+f222/PKaeckj/+4z/OM5/5zJx22ml5whOekA996ENJpu+f+NjHPjYvfvGLc9JJJ+XO\nO+/Me97znhx//PE5/fTT89KXvjTnn39+kuTuu+/O7/7u72b58uVZvnx5PvOZz2Tz5s1517velb/4\ni7/IKaeckn/+53+e5Jc8Mp0uAGBsLrroonz5y1/O9ddfn+3bt+e+++7LL/7iL+Y73/lOzjjjjJx1\n1llJkltvvTXve9/7csYZZ+Qb3/hG3vzmN+fzn/98Dj300Pzmb/7mjlsEvfrVr84f//Ef56lPfWq+\n/vWv57nPfW5uuummvOxlL8shhxyS1772tZP8ch8UoQsA2Cdaa3n961+fT33qU3nIQx6Su+66K9/6\n1reSJMcee2zOOOOMJMnnPve5/MZv/EYe8YhHJEnOOeec3HLLLUmSq6++Ops2bdrxnj/4wQ923AR7\nvhG6AIB9Yu3atbn77rtz3XXX5YADDsiSJUuybdu2JMnBBx880nvcf//9ufbaa3PQQQfty1K7sKYL\nWHCWrL5y1i+gj0MPPTQ//OEPkyTf//7380u/9Es54IADcs011+RrX/va0HOWL1+eT37yk/ne976X\n7du354orrtjx2nOe85y84x3v2PH8+uuvn/XnzBc6XQCwgPW+mveRj3xknvKUp+Skk07K8uXLc/PN\nN+cJT3hCli1blhNOOGHoOUceeWRe//rX5/TTT88jHvGInHDCCTnssMOSJG9/+9vzyle+MieffHK2\nb9+epz/96XnXu96V5z3veXnBC16QD33oQ3nHO96Rpz3taT2/zD0idAEAYzXKDa+//OUv/9zzF73o\nRVm1alW2b9+e3/md38nzn//8JMmjHvWovP/97591/vHHH58bbrhhPAV3YnoRAJi4qampnHLKKTnp\npJOydOnSHaFrIdHpAgAm7uKLL550CfucThcALDCttUmXsOCM4zMVugBgATnooINyzz33CF5j1FrL\nPffcs9fbVpheBIAF5KijjsqWLVty9913T7qUBeWggw7KUUcdtVfvIXQBwAJywAEHZOnSpZMugyFM\nLwIAdDBSp6uqVib5yyT7Jfnr1tpFuzjud5N8MMny1trGwdjrkvxhkp8meVVr7ePjKBxYfIbtLH/e\nyvH/OVPrp2aPdd5gElh4dhu6qmq/JO9M8uwkW5JsqKp1rbVNOx13aJJXJ/nXGWMnJjk3yeOTPDrJ\n1VV1fGvtp+P7EgAA5r5RphdPT3Jba+2O1tqPk1ye5Owhx705yZ8n2TZj7Owkl7fWftRa+2qS2wbv\nBwCwqIwSuo5McueM51sGYztU1WlJjm6t7dz73+25AACLwV4vpK+qhyR5W5L/Zy/eY1VVbayqjS5x\nBQAWolEW0t+V5OgZz48ajP3MoUlOSrK+qpLkf0uyrqrOGuHcJElrbU2SNUmybNkyu7kB3QxbNA+w\nL4zS6dqQ5LiqWlpVB2Z6Yfy6n73YWvt+a+1RrbUlrbUlSa5Nctbg6sV1Sc6tqodW1dIkxyX53Ni/\nCgCAOW63na7W2vaqOj/JxzO9ZcSlrbUbq+qCJBtba+se4Nwbq+oDSTYl2Z7kla5cBPbU1v3XDhk9\nfqRzdbSASRtpn67W2lVJrtpp7I27OHbFTs/fkuQte1gfwAO65OpbZo0dnuUTqATggdmRHgCgA/de\nBBaFYR2xJHnNs0abngTYW0IXsKgNC2OCGLAvmF4EAOhA6AIA6EDoAgDoQOgCAOjAQnpgTrKZKbDQ\n6HQBAHQgdAEAdGB6EZi4JauvnDV23so9f7/h92jcO8OmO6dWzB4D2BWhC2Ane7NhqnAG7IrpRQCA\nDoQuAIAOTC8CjGDYlOPUiv51APOXThcAQAc6XcA+s9AXlY/7qktgYdPpAgDoQKcLmLjh+2qNtkUD\nwHyh0wUA0IHQBQDQgdAFANCBNV3AnDRsXyyA+UzoAvaZhb6hqAsAgAdD6AK6GrZ3F8BiYE0XAEAH\nOl0A+9iwnes3X3TmBCoBJkmnCwCgA6ELAKADoQsAoAOhCwCgA6ELAKADVy8C7GPDN1F19SIsNjpd\nAAAdCF0AAB0IXQAAHVjTBXQ17CbYAIuBThcAQAc6XcCDNrV+atIlzFk6ecCu6HQBAHQgdAEAdCB0\nAQB0IHQBAHQgdAEAdCB0AQB0IHQBAHRgny6ACRi219nUitljwMKh0wUA0IHQBQDQgdAFANCB0AUA\n0IGF9MBYuNEzwAPT6QIA6EDoAgDoYKTpxapameQvk+yX5K9baxft9PrLkrwyyU+T3JtkVWttU1Ut\nSXJTkq8MDr22tfay8ZQOMH8Nm45978eunDW2+aIze5QDdLDb0FVV+yV5Z5JnJ9mSZENVrWutbZpx\n2GWttXcNjj8ryduSrBy8dntr7ZTxlg0AML+MMr14epLbWmt3tNZ+nOTyJGfPPKC19oMZTw9O0sZX\nIgDA/DfK9OKRSe6c8XxLkl/f+aCqemWSP0lyYJLfnPHS0qr6QpIfJHlDa+2fh5y7KsmqJDnmmGNG\nLh5gIdm6/9oho6YXYaEY20L61to7W2v/KcmfJnnDYPibSY5prZ2a6UB2WVX94pBz17TWlrXWlh1x\nxBHjKgkAYM4YJXTdleToGc+PGoztyuVJnp8krbUftdbuGTy+LsntSY7fs1IBAOavUULXhiTHVdXS\nqjowyblJ1s08oKqOm/H0zCS3DsaPGCzET1U9JslxSe4YR+EAAPPJbtd0tda2V9X5ST6e6S0jLm2t\n3VhVFyTZ2Fpbl+T8qnpWkp8k+V6SlwxOf3qSC6rqJ0nuT/Ky1tp398UXAgAwl420T1dr7aokV+00\n9sYZj1+9i/OuSHLF3hQIALAQ2JEeAKADoQsAoAOhCwCgA6ELAKADoQsAoIORrl4EFq+p9VOTLgFg\nQdDpAgDoQOgCAOjA9CLwgC65+pZZY695lluoAjxYOl0AAB3odAHMYUtWXzlrbPNFZ06gEmBv6XQB\nAHQgdAEAdCB0AQB0YE0XwBy2df+1Q0at6YL5SKcLAKADnS7gQRu2dxcAD0ynCwCgA50ugHnG3l0w\nP+l0AQB0oNMF7DC1fmrSJQAsWDpdAAAdCF0AAB0IXQAAHQhdAAAdCF0AAB0IXQAAHQhdAAAdCF0A\nAB0IXQAAHQhdAAAdCF0AAB249yLAPLN1/7VDRs/sXgfw4Oh0AQB0oNMF7HDJ1bdMugSABUunCwCg\nA6ELAKAD04sAC8DU+qnZYytmjwGTI3TBIrVk9ZWzB/0fAWCfMb0IANCBf9cCLFDDupmbL7KfF0yK\nThcAQAdCFwBAB6YXYZEafisZAPYVoQtgARh2N4HDs3wClQC7YnoRAKADoQsAoAOhCwCgA6ELAKAD\noQsAoAOhCwCgA6ELAKADoQsAoIORQldVrayqr1TVbVW1esjrL6uqL1XV9VX16ao6ccZrrxuc95Wq\neu44iwcAmC92uyN9Ve2X5J1Jnp1kS5INVbWutbZpxmGXtdbeNTj+rCRvS7JyEL7OTfL4JI9OcnVV\nHd9a++mYvw7gAUytn5p0CQCL3iidrtOT3NZau6O19uMklyc5e+YBrbUfzHh6cJI2eHx2kstbaz9q\nrX01yW2D9wMAWFRGuffikUnunPF8S5Jf3/mgqnplkj9JcmCS35xx7rU7nXvkHlUKADCPjW0hfWvt\nna21/5TkT5O84cGcW1WrqmpjVW28++67x1USAMCcMUrouivJ0TOeHzUY25XLkzz/wZzbWlvTWlvW\nWlt2xBFHjFASAMD8Mkro2pDkuKpaWlUHZnph/LqZB1TVcTOenpnk1sHjdUnOraqHVtXSJMcl+dze\nlw0AML/sdk1Xa217VZ2f5ONJ9ktyaWvtxqq6IMnG1tq6JOdX1bOS/CTJ95K8ZHDujVX1gSSbkmxP\n8kpXLsJ4LFl95ayxzRedOYFKmE+GXck6tWL2GDB+oyykT2vtqiRX7TT2xhmPX/0A574lyVv2tEAA\ngIXAjvQAAB2M1OkC5rdLrr5l0iUALHo6XQAAHQhdAAAdmF4EWKC27r92yOjx3esApul0AQB0oNMF\nC8iwvbuS+JsOMAfodAEAdCB0AQB0IHQBAHQgdAEAdCB0AQB04JomgEVk2C2hplb0rwMWI50uAIAO\nhC4AgA6ELgCADoQuAIAOhC4AgA6ELgCADmwZAQvI1v3XTroEAHZBpwsAoAOhCwCgA6ELAKADoQsA\noAOhCwCgA6ELAKADW0YALHJT66eGj68YPg7sGZ0uAIAOhC4AgA6ELgCADqzpAljkLrn6lqHjUyv6\n1gELnU4XAEAHQhcAQAemF2Ge2rr/2kmXwAI3bCsJ20jAntPpAgDoQKcLgKGGLbC3uB72nNAF88Cu\ndgwHYP4wvQgA0IHQBQDQgdAFANCB0AUA0IHQBQDQgdAFANCB0AUA0IF9umAeGLZJJQDzi04XAEAH\nQhcAQAemF2GOWbL6ytmD/qYCzHs6XQAAHQhdAAAdCF0AAB0IXQAAHQhdAAAdCF0AAB2MFLqqamVV\nfaWqbquq1UNe/5Oq2lRVN1TV/1dVx8547adVdf3g17pxFg8AMF/sdvefqtovyTuTPDvJliQbqmpd\na23TjMO+kGRZa+2+qnp5kv+a5PcGr/1Ha+2UMdcNADCvjNLpOj3Jba21O1prP05yeZKzZx7QWrum\ntXbf4Om1SY4ab5kAAPPbKKHryCR3zni+ZTC2K3+Y5KMznh9UVRur6tqqev6wE6pq1eCYjXffffcI\nJQEAzC9jvblIVf1BkmVJfmPG8LGttbuq6jFJ/qmqvtRau33mea21NUnWJMmyZcvaOGsCAJgLRul0\n3ZXk6BnPjxqM/ZyqelaSP0tyVmvtRz8bb63dNfj9jiTrk5y6F/UCAMxLo4SuDUmOq6qlVXVgknOT\n/NxViFV1apJ3ZzpwfXvG+MOr6qGDx49K8pQkMxfgAwAsCrudXmytba+q85N8PMl+SS5trd1YVRck\n2dhaW5fkrUkOSfJ3VZUkX2+tnZXkcUneXVX3ZzrgXbTTVY8AAIvCSGu6WmtXJblqp7E3znj8rF2c\n99kkT9ibAgEAFgI70gMAdCB0AQB0MNYtI4C9t3X/tZMuAXbp8De8aNbY1gsvm0AlMP/odAEAdCB0\nAQB0IHQBAHQgdAEAdCB0AQB0IHQBAHQgdAEAdCB0AQB0IHQBAHQgdAEAdCB0AQB0IHQBAHQgdAEA\ndLD/pAuAxWxq/dSkSwCgE50uAIAOhC4AgA6ELgCADoQuAIAOhC4AgA6ELgCADmwZARN0ydW3TLoE\nADrR6QIA6EDoAgDoQOgCAOhA6AIA6MBCegD2yrB7iE6tmD0Gi51OFwBABzpdAOyVYVufTK3oXwfM\ndTpdAAAd6HQBMHbWecFsOl0AAB0IXQAAHQhdAAAdCF0AAB1YSA+dLFl95exBfwMBFg3/ywdg7Ozd\nBbOZXgQA6EDoAgDoQOgCAOhA6AIA6MBCetgHhl6pCMCiptMFANCB0AUA0IHpRehk6/5rJ10CABOk\n0wUA0IHQBQDQgdAFANCB0AUA0IHQBQDQgdAFANDBSFtGVNXKJH+ZZL8kf91au2in1/8kyf+VZHuS\nu5P859ba1wavvSTJGwaHXthae9+Yaoc5y/YQMNvU+qnZYytmj8FCtdtOV1Xtl+SdSX4ryYlJXlhV\nJ+502BeSLGutnZzkg0n+6+DcRyR5U5JfT3J6kjdV1cPHVz4AwPwwyvTi6Ulua63d0Vr7cZLLk5w9\n84DW2jWttfsGT69NctTg8XOT/GNr7butte8l+cckK8dTOgDA/DFK6DoyyZ0znm8ZjO3KHyb56IM5\nt6pWVdXGqtp49913j1ASAMD8MtaF9FX1B0mWJXnrgzmvtbamtbastbbsiCOOGGdJAABzwiih664k\nR894ftRg7OdU1bOS/FmSs1prP3ow5wIALHSjhK4NSY6rqqVVdWCSc5Osm3lAVZ2a5N2ZDlzfnvHS\nx5M8p6oePlhA/5zBGADAorLbLSNaa9ur6vxMh6X9klzaWruxqi5IsrG1ti7T04mHJPm7qkqSr7fW\nzmqtfbeq3pzp4JYkF7TWvrtPvhIAgDlspH26WmtXJblqp7E3znj8rAc499Ikl+5pgQAAC4Ed6QEA\nOhC6AAA6GGl6EQD21iVX3zJrbGpF/zpgUoQu2EtLVl85e9DfLBiJ+zGymPjRAA/C0IAFACOwpgsA\noAOhCwCgA6ELAKADoQsAoAOhCwCgA6ELAKADoQsAoAOhCwCgA5ujwl7auv/aSZcAwDwgdAEwMe7H\nyGJiehEAoAOhCwCgA6ELAKADoQsAoAOhCwCgA6ELAKADW0YAMKdMrZ+aPbZi9hjMN0IX7MKS1VdO\nugQAFhDTiwAAHQhdAAAdmF6EB8F9FgHYUzpdAAAdCF0AAB0IXQAAHQhdAAAdCF0AAB0IXQAAHQhd\nAAAdCF0AAB0IXQAAHQhdAAAdCF0AAB0IXQAAHbjhNQBzyiVX3zJrbGpF/zpg3IQu2IWt+6+ddAkA\nLCCmFwEAOhC6AAA6ELoAADoQugAAOhC6AAA6ELoAADoQugAAOhC6AAA6sDkqAHPe1Pqp2WMrZo/B\nXKbTBQDQgU4XAHOe+zGyEAhdkOFTFwAwTqYXAQA6GKnTVVUrk/xlkv2S/HVr7aKdXn96kkuSnJzk\n3NbaB2e89tMkXxo8/Xpr7axxFA7jNGzqAgDGabehq6r2S/LOJM9OsiXJhqpa11rbNOOwryc5L8lr\nh7zFf7TWThlDrQAA89Yona7Tk9zWWrsjSarq8iRnJ9kRulprmwev3b8PagQAmPdGWdN1ZJI7Zzzf\nMhgb1UFVtbGqrq2q5z+o6gAAFogeVy8e21q7q6oek+SfqupLrbXbZx5QVauSrEqSY445pkNJAAB9\njdLpuivJ0TOeHzUYG0lr7a7B73ckWZ/k1CHHrGmtLWutLTviiCNGfWsAgHljlNC1IclxVbW0qg5M\ncm6SdaO8eVU9vKoeOnj8qCRPyYy1YAAAi8VuQ1drbXuS85N8PMlNST7QWruxqi6oqrOSpKqWV9WW\nJOckeXdV3Tg4/XFJNlbVF5Nck+Sina56BABYFEZa09VauyrJVTuNvXHG4w2Znnbc+bzPJnnCXtYI\nADDv2ZEeAKADoQsAoAM3vAZgXlqy+spZY5svOnMClcBohC4AFgxBjLnM9CIAQAdCFwBAB0IXAEAH\n1nQBMC9t3X/trLHDt//+BCqB0eh0AQB0oNPFojPs6iZ/EwDY13S6AAA6ELoAADoQugAAOrCSBYAF\nY9gVjYkd6ZkbdLoAADoQugAAOjC9yKIzfPoBAPYtnS4AgA6ELgCADoQuAIAOrOkCYEGbWj81e2zF\n7DHY13S6AAA6ELoAADoQugAAOhC6AAA6ELoAADoQugAAOhC6AAA6ELoAADoQugAAOhC6AAA6cBsg\nFrQlq6+cPei7HoAJ0OkCAOjAv/lZ0Lbuv3bSJQATdsnVt8wam1rRvw7Q6QIA6EDoAgDoQOgCAOhA\n6AIA6EDoAgDowNWLLAhD9+NKfIcDMGfodAEAdCB0AQB0YPKFBcEmqADMdTpdAAAd6HQBsOgMu/hm\n80VnTqASFhOhC4BFZ/iSBKGLfcv0IgBAB0IXAEAHpheZd4ZuhOo7GYA5TqcLAKADoQsAoAOhCwCg\nAythACDJ1Pqp2WMrZo/BntLpAgDoYKTQVVUrq+orVXVbVa0e8vrTq+rzVbW9ql6w02svqapbB79e\nMq7CWby27r921i8AmOt2G7qqar8k70zyW0lOTPLCqjpxp8O+nuS8JJftdO4jkrwpya8nOT3Jm6rq\n4XtfNgDA/DJKp+v0JLe11u5orf04yeVJzp55QGttc2vthiT373Tuc5P8Y2vtu6217yX5xyQrx1A3\nAMC8MspC+iOT3Dnj+ZZMd65GMezcI3c+qKpWJVmVJMccc8yIb81iMGxhKwDMR3Pi6sXW2poka5Jk\n2bJlbcLlMIdccvUtky4BAMZilOnFu5IcPeP5UYOxUezNuQAAC8YooWtDkuOqamlVHZjk3CTrRnz/\njyd5TlU9fLCA/jmDMQCARWW304utte1VdX6mw9J+SS5trd1YVRck2dhaW1dVy5P8Q5KHJ3leVf2X\n1trjW2vfrao3Zzq4JckFrbXv7qOvBQD22LDlDFMr+tfBwjXSmq7W2lVJrtpp7I0zHm/I9NThsHMv\nTXLpXtQIADDv2ZEeAKADoQsAoAOhCwCgA6ELAKADoQsAoAOhCwCgA6ELAKCDOXHvRQCYi6bWTw0f\nXzF8HB6IThcAQAdCFwBAB6YXAWAXht2PMXFPRvaMThcAQAc6XcwZS1ZfOXvQdygwBw1bYG9xPbuj\n0wUA0IHQBQDQgckbABiDYUskNl905gQqYa7S6QIA6EDoAgDoQOgCAOhA6AIA6MBCeuaMrfuvnXQJ\nACMZtlP94Vk+gUqYT3S6AAA6ELoAADoQugAAOhC6AAA6sJCeiXBzawAWGz/mmAhXKgKw2JheBADo\nQKcLAMZgeAffDa/5X4Qu9jnrtwDA9CIAQBdCFwBAB0IXAEAHQhcAQAdCFwBAB64hA4B9ZNjV25sv\nso3EYqXTBQDQgdAFANCB0AUA0IHQBQDQgdAFANCBqxcZq2FX6gy/CSzAwucm2Myk0wUA0IHQBQDQ\ngelFxspUIgAMp9MFANCBThcAdDTsgqPzVm6YNTa1YqpDNfSk0wUA0IFOF3ts2L/WfEcBPLDha1+P\n714H/el0AQB0oC/BHnOlIgCMTqcLAKADoQsAoIORQldVrayqr1TVbVW1esjrD62q9w9e/9eqWjIY\nX1JV/1E7LFRCAAALbElEQVRV1w9+vWu85QMAzA+7XdNVVfsleWeSZyfZkmRDVa1rrW2acdgfJvle\na+1Xq+rcJH+e5PcGr93eWjtlzHUDwIJxydW3zBqbWtG/DvatUTpdpye5rbV2R2vtx0kuT3L2Tsec\nneR9g8cfTPLMqqrxlQkAML+NErqOTHLnjOdbBmNDj2mtbU/y/SSPHLy2tKq+UFWfrKqnDfsDqmpV\nVW2sqo133333g/oCAADmg329ZcQ3kxzTWrunqn4tyf+sqse31n4w86DW2poka5Jk2bJlbR/XxB6w\nESoA7J1ROl13JTl6xvOjBmNDj6mq/ZMcluSe1tqPWmv3JElr7bokt8e2uwDAIjRK6NqQ5LiqWlpV\nByY5N8m6nY5Zl+Qlg8cvSPJPrbVWVUcMFuKnqh6T5Lgkd4yndACA+WO3E0Stte1VdX6SjyfZL8ml\nrbUbq+qCJBtba+uSvCfJ/6iq25J8N9PBLEmenuSCqvpJkvuTvKy19t198YUAwEIytX5q9tiK2WPM\nHyOtymmtXZXkqp3G3jjj8bYk5ww574okV+xljcwBbvkDAHvHjvQAAB0IXQAAHbjoHwDmoGG71L/3\nY7O379l80Zk9ymEMhK5Fbtj+W/4CA8D4mV4EAOhApwsA5onhV5KbnZgvdLoAADrQ6WKWw9/wokmX\nAAALjk4XAEAHQhcAQAdCFwBAB9Z0AcA85sbY84dOFwBABzpdi9zwPV8AgHETugBgHht2j8apFf3r\nYPeELgBYYKzzmpuErkVk2M2tfQcAQB9+5ALAAmPKcW4SuhaoYV0ti+YBYHKErgVKwAKAucU+XQAA\nHQhdAAAdCF0AAB1Y0wUAi8CwC6w2X3TmBCpZvIQuAFgEhl9gJXT1JHQtAMN2HgYA5hahCwAWqcPf\n8KJZY1svvGwClSwOQtcCMGznYQBgbhG65pFh/yIBAOYHW0YAAHQgdAEAdGB6cY4atp+K/1oAMH/p\ndAEAdKB3AgDs8GD2fpxaMfqxCF1z1vCdgwFg3xq2DdFrnnX8BCpZeISuztz7CgAWJ6ELAHhAu9qE\ne2pF3zrmO6FrDnDvRADmI7M3D47QNQe4jQ8ALHy2jAAA6ECnCwDYI8OutF+yevZxphynCV2d2QoC\ngIVs+M85oSsRuvapw9/wokmXAADMEdZ0AQB0oNMFAOxTw2Z+tl542QQqmSyhaw8M25dkKJ8uADAg\nFoyJBfIAMLphG4Mv9BtoC10AQHfDNwafmj2ygIKY0LUbw6YSdbUAYPyGBbGFdH9HoWuGYQv9Ds/v\nT6ASACAZ/f7E86EjJnTthq4WAEzOsO7Xa551/AQq2XuLNnTZuBQA5qf5uh5s0YYuAGDhmA/rwUba\nkb6qVlbVV6rqtqqadSvLqnpoVb1/8Pq/VtWSGa+9bjD+lap67vhKBwCYP3YbuqpqvyTvTPJbSU5M\n8sKqOnGnw/4wyfdaa7+a5C+S/Png3BOTnJvk8UlWJvmrwfsBACwqo3S6Tk9yW2vtjtbaj5NcnuTs\nnY45O8n7Bo8/mOSZVVWD8ctbaz9qrX01yW2D9wMAWFRGCV1HJrlzxvMtg7Ghx7TWtif5fpJHjngu\nAMCCNycW0lfVqiSrBk/vraqvdPhjH5XkOx3+nMXC5zl+PtPx8nmOn890/HymY1Rv+dsen+exox44\nSui6K8nRM54fNRgbdsyWqto/yWFJ7hnx3LTW1iRZM2rR41BVG1try3r+mQuZz3P8fKbj5fMcP5/p\n+PlMx2uufZ6jTC9uSHJcVS2tqgMzvTB+3U7HrEvyksHjFyT5p9ZaG4yfO7i6cWmS45J8bjylAwDM\nH7vtdLXWtlfV+Uk+nmS/JJe21m6sqguSbGytrUvyniT/o6puS/LdTAezDI77QJJNSbYneWVr7af7\n6GsBAJizRlrT1Vq7KslVO429ccbjbUnO2cW5b0nylr2ocV/pOp25CPg8x89nOl4+z/HzmY6fz3S8\n5tTnWdOzgAAA7Esj7UgPAMDeWdShq6reXFU3VNX1VfWJqnr0pGuaz6rqrVV18+Az/YeqOnzSNc13\nVXVOVd1YVfdX1Zy5Ame+2d2tzHhwqurSqvp2VX150rUsBFV1dFVdU1WbBn/fXz3pmua7qjqoqj5X\nVV8cfKb/ZdI1JYt8erGqfrG19oPB41clObG19rIJlzVvVdVzMn3l6vaq+vMkaa396YTLmteq6nFJ\n7k/y7iSvba1tnHBJ887g1mO3JHl2pjdo3pDkha21TRMtbB6rqqcnuTfJf2+tnTTpeua7qvqVJL/S\nWvt8VR2a5Lokz/c9uucGd8U5uLV2b1UdkOTTSV7dWrt2knUt6k7XzwLXwMFJFm8CHYPW2icGdyRI\nkmszvS8be6G1dlNrrcdmwQvZKLcy40ForX0q01eqMwattW+21j4/ePzDJDfF3Vv2Spt27+DpAYNf\nE/8Zv6hDV5JU1Vuq6s4kv5/kjbs7npH95yQfnXQRELcjYx6pqiVJTk3yr5OtZP6rqv2q6vok307y\nj621iX+mCz50VdXVVfXlIb/OTpLW2p+11o5OsjbJ+ZOtdu7b3ec5OObPMr0v29rJVTp/jPKZAgtf\nVR2S5Iokr9lpJoY90Fr7aWvtlEzPupxeVROfCp8T917cl1przxrx0LWZ3ovsTfuwnHlvd59nVZ2X\n5H9P8sy2mBcMPggP4nuUPTPS7chgkgbrjq5Isra19veTrmchaa1traprkqxMMtGLPxZ8p+uBVNVx\nM56eneTmSdWyEFTVyiT/b5KzWmv3TboeGBjlVmYwMYNF3+9JclNr7W2TrmchqKojfnYFfVX9QqYv\npJn4z/jFfvXiFUkem+mrw76W5GWtNf8C3kOD20A9NNM3O0+Sa10Nuneq6neSvCPJEUm2Jrm+tfbc\nyVY1/1TVbye5JP/rVmZz8S4Z80ZV/W2SFUkeleRbSd7UWnvPRIuax6rqqUn+OcmXMv3zKEleP7gb\nDHugqk5O8r5M/51/SJIPtNYumGxVizx0AQD0sqinFwEAehG6AAA6ELoAADoQugAAOhC6AAA6ELqA\nOaeqflpV18/4tXow/pqqethuzj27qv7njOevG2xn8rPnz6uqdYPHV/1sL5+d3mOqql47eHxeVT16\nxmubq+pRe/9VAovNgt+RHpiX/mNw+46dvSbJ3yR5oM13P5vk3TOePynJD6rql1pr307y5MExaa39\n9gi1nJfpXay/McKxALuk0wXMC1X1qiSPTnLN4JYeqar/VlUbq+rGqvovSdJauzvTIetXB6cemenb\nqzx58PzJST4zOH9H16qq/qyqbqmqT2d60+RU1QuSLEuydtBx+4XBe/zfVfX5qvpSVZ2wr792YGEQ\nuoC56Bd2ml78vdba2zPdbXpGa+0Zg+P+rLW2LMnJSX5jsAt1Mh2qnlxVj01ya5JrB8/3T/LETN8a\naIeq+rVM3x7olCS/nWR5krTWPphkY5Lfb62d0lr7j8Ep32mtnZbkvyV57T75BIAFx/QiMBftanpx\nZ/9nVa3K9P/LfiXJiUluyPT04ZMzfQuQf0nyuSRvTHJqkptba9t2ep+nJfmHn90z9Gdrvh7Az25I\nfF2S/2OEOgF0uoD5qaqWZrrL9MzW2slJrkxy0ODlz2Q6dD05yb+01n44eG1FBuu59tKPBr//NP7x\nCoxI6ALmkx8mOXTw+BeT/HuS71fVLyf5rRnH3ZTp9V9PTfKFwdj1SV6WwXqunXwqyfOr6heq6tAk\nz9vFnwmwx/wLDZiLfqGqrp/x/GOttdVJ1iT5WFV9o7X2jKr6QpKbk9yZGWGqtdaq6l+THNZa+8lg\n+F+SrMqQTldr7fNV9f4kX0zy7fz8mq/3JnlXVf1Hpq+EBNgj1VqbdA0AAAue6UUAgA6ELgCADoQu\nAIAOhC4AgA6ELgCADoQuAIAOhC4AgA6ELgCADv5/ezzXrbHdN4wAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2c3f43ee10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhC4AgA6ELgCADv5/GX5yqionvCwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6bbafd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6c6d898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bfce1d550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAHQgdAEAdCB0AQB0IHQBAHTw/wG1W5WhcOIFygAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bfcddccf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6c68860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAJQCAYAAAC0HNH1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xu0nWV9L/rvD4KwBQqKtKdyS2xBREDAhOINY0VID1vQ\nUzmidStnd5h64Xjpce+dWoeuIuxDK8emeujW7Mqopw3ihbZGwUvTEq+lJLGIEjBcTCVoFdGolAYI\nPOePNUmXWTNkJpl557p8PmOskfk+72X95mQl68vzPO/zVmstAADsWXuNugAAgNlA6AIA6IDQBQDQ\nAaELAKADQhcAQAeELgCADghdAAAdELoAADogdAEAdGDOqAvY1pOe9KQ2d+7cUZcBALBDa9eu/WFr\n7dBBjp1yoWvu3LlZs2bNqMsAANihqvrnQY81vAgA0AGhCwCgA0IXAEAHptycLgBg1z300EPZuHFj\nNm/ePOpSZpT99tsvhx9+ePbZZ59dvobQBQAzyMaNG3PggQdm7ty5qapRlzMjtNZy7733ZuPGjZk3\nb94uX8fwIgDMIJs3b84hhxwicA1RVeWQQw7Z7d5DoQsAZhiBa/iG8ZkKXQAAHTCnCwBmsLlLrhnq\n9TZcevZQr7cnLV26NIsXL87jH//4gc9ZtWpVLrvssnz6058eej16ugCAaam1lkceeWS7+5cuXZr7\n77+/w4oem9AFAAzVu9/97jz1qU/Nc5/73LziFa/IZZddljvuuCOLFi3KM5/5zDzvec/LrbfemiS5\n4IIL8qY3vSnPfvaz85SnPCWf+MQntl7nPe95TxYsWJATTzwx73rXu5IkGzZsyFOf+tS8+tWvzvHH\nH5+77rorr3/96zN//vw8/elP33rc+973vnz3u9/NC17wgrzgBS9Iknz+85/Ps571rJxyyik577zz\nct999yVJPvvZz+bYY4/NKaeckr/6q7/aY5+L0AUADM3q1atz9dVX5+tf/3o+85nPbH2e8uLFi/P+\n978/a9euzWWXXZY3vOENW8/53ve+ly9/+cv59Kc/nSVLliQZD0i33XZbbrjhhtx4441Zu3ZtvvjF\nLyZJbrvttrzhDW/IzTffnKOOOiqXXHJJ1qxZk5tuuilf+MIXctNNN+VNb3pTnvzkJ+e6667Ldddd\nlx/+8Ie5+OKLs3Llynzta1/L/Pnz8973vjebN2/Oa1/72nzqU5/K2rVr8y//8i977LMxpwsAGJqv\nfOUrOffcc7Pffvtlv/32y4tf/OJs3rw5X/3qV3PeeedtPe6BBx7Y+volL3lJ9tprrxx33HH5/ve/\nn2Q8dH3+85/PySefnCS57777ctttt+XII4/MUUcdldNOO23r+R/72MeybNmybNmyJd/73veybt26\nnHjiiT9X1/XXX59169blOc95TpLkwQcfzLOe9azceuutmTdvXo4++ugkyate9aosW7Zsj3w2QhcA\nsEc98sgjOfjgg3PjjTf23b/vvvtufd1a2/rn7/3e7+V3fud3fu7YDRs2ZP/999+6/e1vfzuXXXZZ\nVq9enSc84Qm54IIL+q6n1VrLi170onzkIx/5ufbt1bQnGF4EAIbmOc95Tj71qU9l8+bNue+++/Lp\nT386j3/84zNv3rx8/OMfTzIegL7+9a8/5nXOOuusXHHFFVvnXd199935wQ9+MOm4n/70p9l///1z\n0EEH5fvf/34+85nPbN134IEH5mc/+1mS5LTTTstXvvKV3H777UmSf/3Xf8369etz7LHHZsOGDbnj\njjuSZFIoGyY9XQAwg3W9xMOCBQtyzjnn5MQTT8wv/dIv5YQTTshBBx2U5cuX5/Wvf30uvvjiPPTQ\nQzn//PPzjGc8Y7vXOfPMM3PLLbfkWc96VpLkgAMOyF/+5V9m7733/rnjnvGMZ+Tkk0/OsccemyOO\nOGLr8GEyPo9s0aJFW+d2/fmf/3le8YpXbB3avPjii3PMMcdk2bJlOfvss/P4xz8+z3ve87YGtWGr\nR7vxpor58+e3RyfdAQA755ZbbsnTnva0kdZw33335YADDsj999+f008/PcuWLcspp5wy0pqGod9n\nW1VrW2vzBzlfTxcAMFSLFy/OunXrsnnz5rzmNa+ZEYFrGIQuAGCorrzyylGXMCWZSA8A0AGhCwCg\nA0IXAEAHhC4AgA6YSA8AM9jYqrHhXm/hY19v06ZNufLKK3/u2Yp7wqpVq/K4xz0uz372s/fo9xkm\noQuYFeYuuaZve9cLR8JMt2nTpvzpn/7pwKGrtZbWWvbaa+cG31atWpUDDjhgWoUuw4sAwNAsWbIk\nd9xxR0466aS89a1vzQtf+MKccsopOeGEE/LJT34yyfjzE5/61Kfm1a9+dY4//vjcdddd+dCHPpRj\njjkmp556al772tfmwgsvTJLcc889+c3f/M0sWLAgCxYsyFe+8pVs2LAhH/jAB/LHf/zHOemkk/Kl\nL31plG95YHq6AIChufTSS/PNb34zN954Y7Zs2ZL7778/v/ALv5Af/vCHOe2003LOOeckSW677bZ8\n+MMfzmmnnZbvfve7efe7352vfe1rOfDAA/Prv/7rWx8R9OY3vzlvfetb89znPjff+c53ctZZZ+WW\nW27J6173uhxwwAF529veNsq3u1OELmBa6zdsaMgQpobWWt7+9rfni1/8Yvbaa6/cfffd+f73v58k\nOeqoo3LaaaclSW644YY8//nPzxOf+MQkyXnnnZf169cnSVauXJl169ZtveZPf/rTrQ/Bnm6ELgBg\nj1i+fHnuueeerF27Nvvss0/mzp2bzZs3J0n233//ga7xyCOP5Prrr89+++23J0vthDldAMDQHHjg\ngfnZz36WJPnJT36SX/zFX8w+++yT6667Lv/8z//c95wFCxbkC1/4Qn784x9ny5Ytufrqq7fuO/PM\nM/P+979/6/aNN9446ftMF3q6AGAG29ESD8N2yCGH5DnPeU6OP/74LFiwILfeemtOOOGEzJ8/P8ce\ne2zfcw477LC8/e1vz6mnnponPvGJOfbYY3PQQQclSd73vvfljW98Y0488cRs2bIlp59+ej7wgQ/k\nxS9+cV72spflk5/8ZN7//vfnec97Xpdvc5cIXQDAUA3ywOtvfvObP7f9yle+MosXL86WLVvy0pe+\nNC95yUuSJE960pPy0Y9+dNL5xxxzTG666abhFNwRw4sAwMiNjY3lpJNOyvHHH5958+ZtDV0ziZ4u\nAGDkLrvsslGXsMcJXcCs1u8RKV3PgYFha62lqkZdxozSWtvtawhdwKywac7y7ew5ZqDzD37HKydf\n8+Idz1uBru2333659957c8ghhwheQ9Jay7333rvby1YMFLqqalGSP0myd5I/a61dup3jfjPJJ5Is\naK2t6bX9XpLfTvJwkje11j63WxUDANt1+OGHZ+PGjbnnnntGXcqMst9+++Xwww/frWvsMHRV1d5J\nLk/yoiQbk6yuqhWttXXbHHdgkjcn+ccJbcclOT/J05M8OcnKqjqmtfbwblUNAPS1zz77ZN68eaMu\ngz4G6ek6NcntrbU7k6SqrkpybpJ12xz37iR/mOS/TGg7N8lVrbUHkny7qm7vXe8fdrdwgFEzHwzY\nGYMsGXFYkrsmbG/stW1VVackOaK1tu1D0HZ4LgDAbLDb63RV1V5J3pvk/9qNayyuqjVVtcYYNAAw\nEw0Suu5OcsSE7cN7bY86MMnxSVZV1YYkpyVZUVXzBzg3SdJaW9Zam99am3/ooYfu3DsAAJgGBgld\nq5McXVXzqupxGZ8Yv+LRna21n7TWntRam9tam5vk+iTn9O5eXJHk/Krat6rmJTk6yQ1DfxcAAFPc\nDifSt9a2VNWFST6X8SUjrmit3VxVFyVZ01pb8Rjn3lxVH8v4pPstSd7ozkVgKlm6cv2ktj//7LbT\nU2NVQ2C3DfTPSGvt2iTXbtP2zu0cu3Cb7UuSXLKL9QEAzAj+3w2YknZnOYZ+5wKMmtAFTBtzl/QZ\n9gOYJoQuYMbpN08LYNR2e50uAAB2TE8XMCX16606OAs6+d6b5izf5XM9GgjYHqELYBf1C4ZvOeOY\nEVQCTAeGFwEAOqCnCxi5QZd46Dfsd/CW3xpyNQB7hp4uAIAOCF0AAB0wvAiM3O6sq7U7dxoCdElP\nFwBAB4QuAIAOCF0AAB0QugAAOiB0AQB0QOgCAOiA0AUA0AHrdAHsYf0eczS2cHIbMLPp6QIA6IDQ\nBQDQAcOLQKfmLrlmcqN/iYBZwD91wB7TN2ABzFKGFwEAOiB0AQB0wPAiwBAtXbl+UttbzjhmBJUA\nU43QBXRq05zloy4BYCQMLwIAdEBPFzAU7lQEeGx6ugAAOiB0AQB0QOgCAOiA0AUA0AGhCwCgA+5e\nBIbC+lsAj01PFwBAB/R0Aexh/R4NNLaw+zqA0dLTBQDQAaELAKADQhcAQAeELgCADghdAAAdELoA\nADpgyQiAKWLukmsmtW249OwRVALsCXq6AAA6IHQBAHRA6AIA6IDQBQDQAaELAKADQhcAQAeELgCA\nDghdAAAdsDgqwAiMrRrr07qg6zKADunpAgDowEA9XVW1KMmfJNk7yZ+11i7dZv/rkrwxycNJ7kuy\nuLW2rqrmJrklybd6h17fWnvdcEoHRqXf42r0mwM8th3+M1lVeye5PMmLkmxMsrqqVrTW1k047MrW\n2gd6x5+T5L1JFvX23dFaO2m4ZQMATC+DDC+emuT21tqdrbUHk1yV5NyJB7TWfjphc/8kbXglAgBM\nf4MMCByW5K4J2xuT/Nq2B1XVG5P8bpLHJfn1CbvmVdU/Jflpkne01r606+UCzAxLV66f1HawifQw\now1tIn1r7fLW2q8k+W9J3tFr/l6SI1trJ2c8kF1ZVb+w7blVtbiq1lTVmnvuuWdYJQEATBmD9HTd\nneSICduH99q256ok/yNJWmsPJHmg93ptVd2R5Jgkayae0FpblmRZksyfP9/QJDArbZqzvE/r2Z3X\nAewZg/R0rU5ydFXNq6rHJTk/yYqJB1TV0RM2z05yW6/90N5E/FTVU5IcneTOYRQOADCd7LCnq7W2\npaouTPK5jC8ZcUVr7eaquijJmtbaiiQXVtUZSR5K8uMkr+mdfnqSi6rqoSSPJHlda+1He+KNAABM\nZQOtrNNauzbJtdu0vXPC6zdv57yrk1y9OwUCAMwEVqQHAOiA0AUA0AGhCwCgA0IXAEAHhC4AgA4I\nXQAAHRC6AAA6MNA6XcDsNbZqbFLbpjmTH9YMwGPT0wUA0AGhCwCgA0IXAEAHhC4AgA6YSA88pqUr\nTZofpblLrpnUdsGi1ZPaxhaOdVANsDv0dAEAdEBPF8AUtmnO8j6tx3ReB7D79HQBAHRA6AIA6IDQ\nBQDQAaELAKADJtIDW/VbnsC/EgDDoacLAKADQhcAQAcMHABMM/2eEjC2sPs6gJ2jpwsAoANCFwBA\nB4QuAIAOCF0AAB0QugAAOiB0AQB0QOgCAOiA0AUA0AGhCwCgA0IXAEAHPAYI2GrTnOWjLgFgxtLT\nBQDQAaELAKADQhcAQAfM6YJZamzV2KhLAJhVhC6YpZauXD/qEgBmFaELYAaYu+SaSW0bLj17BJUA\n2yN0AcwA/Zf7ELpgKjGRHgCgA0IXAEAHhC4AgA4IXQAAHRC6AAA6IHQBAHRA6AIA6IDQBQDQAaEL\nAKADQhcAQAeELgCADghdAAAdELoAADowUOiqqkVV9a2qur2qlvTZ/7qq+kZV3VhVX66q4ybs+73e\ned+qqrOGWTwAwHSxw9BVVXsnuTzJbyQ5LskrJoaqnitbaye01k5K8kdJ3ts797gk5yd5epJFSf60\ndz0AgFllkJ6uU5Pc3lq7s7X2YJKrkpw78YDW2k8nbO6fpPVen5vkqtbaA621bye5vXc9AIBZZc4A\nxxyW5K4J2xuT/Nq2B1XVG5P8bpLHJfn1Cedev825h/U5d3GSxUly5JFHDlI3AMC0MkjoGkhr7fIk\nl1fVK5O8I8lrduLcZUmWJcn8+fPbDg4HdtLYqrFRlwAw6w0Suu5OcsSE7cN7bdtzVZL/sYvnAjAk\nc5dcM6ltw6Vnj6ASIBksdK1OcnRVzct4YDo/ySsnHlBVR7fWbuttnp3k0dcrklxZVe9N8uQkRye5\nYRiFA4NbunL9qEsAmPV2GLpaa1uq6sIkn0uyd5IrWms3V9VFSda01lYkubCqzkjyUJIfpze02Dvu\nY0nWJdmS5I2ttYf30HsBYIJNc5b3adXTBaMy0Jyu1tq1Sa7dpu2dE16/+THOvSTJJbtaIADATGBF\negCADgzt7kUApr5+d7KOLZzcBgyfni4AgA4IXQAAHRC6AAA6IHQBAHRA6AIA6IDQBQDQAaELAKAD\nQhcAQAeELgCADghdAAAdELoAADogdAEAdEDoAgDowJxRFwAMz9iqsVGXAMB26OkCAOiA0AUA0AHD\niwCzyNKV6ye1jS3svg6YjfR0AQB0QE8XTFNzl1wzqe2CRSMoBICBCF0wg/QbOgJgahC6YJraNGf5\nqEsAYCeY0wUA0AGhCwCgA0IXAEAHhC4AgA4IXQAAHRC6AAA6YMkIgFmu30K7SbLh0rM7rgRmNj1d\nAAAd0NMFMMttf6FdPV0wTHq6AAA6IHQBAHRA6AIA6IDQBQDQAaELAKADQhcAQAeELgCADghdAAAd\nELoAADogdAEAdEDoAgDogGcvAtDX2KqxyW0LJ7cBg9HTBQDQAT1dMA3MXXLN5EZ/ewGmFT1dAAAd\nELoAADpggAKmgU1zlo+6BAB2k9AFQF9LV66f1Da2sPs6YKYwvAgA0AGhCwCgA0IXAEAHhC4AgA4M\nFLqqalFVfauqbq+qJX32/25Vrauqm6rq76rqqAn7Hq6qG3tfK4ZZPADAdLHDuxerau8klyd5UZKN\nSVZX1YrW2roJh/1Tkvmttfur6vVJ/ijJy3v7/q21dtKQ64YZy+rzADPTID1dpya5vbV2Z2vtwSRX\nJTl34gGttetaa/f3Nq9PcvhwywQAmN4GCV2HJblrwvbGXtv2/HaSz0zY3q+q1lTV9VX1kn4nVNXi\n3jFr7rnnngFKAgCYXoY6aFFVr0oyP8nzJzQf1Vq7u6qekuTvq+obrbU7Jp7XWluWZFmSzJ8/vw2z\nJgCAqWCQnq67kxwxYfvwXtvPqaozkvx+knNaaw882t5au7v3551JViU5eTfqBQCYlgYJXauTHF1V\n86rqcUnOT/JzdyFW1clJPpjxwPWDCe1PqKp9e6+flOQ5SSZOwAcAmBV2OLzYWttSVRcm+VySvZNc\n0Vq7uaouSrKmtbYiyXuSHJDk41WVJN9prZ2T5GlJPlhVj2Q84F26zV2PAACzwkBzulpr1ya5dpu2\nd054fcZ2zvtqkhN2p0AAgJnAivQAAB0QugAAOmCdawAGNrZqbHLbwsltwGR6ugAAOiB0AQB0QOgC\nAOiA0AUA0AGhCwCgA0IXAEAHhC4AgA5YpwuA3TJ3yTWT2jZcevYIKoGpTeiCKWbTnOWjLgGAPcDw\nIgBAB4QuAIAOCF0AAB0QugAAOmAiPQADW7py/aS2g7NgBJXA9KOnCwCgA3q6ANgt/Zc5sU4XbEtP\nFwBAB4QuAIAOCF0AAB0QugAAOmAiPYzQ2KqxUZcAQEeELgCGbu6Saya1bbjUHY3MboYXAQA6IHQB\nAHRA6AIA6IA5XTBC/Z5jB8DMJHRBR/pNLPY3EGD2MLwIANABoQsAoANCFwBAB4QuAIAOCF0AAB1w\n7xQAQ7dpzvI+rR4DxOwmdEFH+v8SAmC2MLwIANABoQsAoANCFwBAB4QuAIAOCF0AAB0QugAAOiB0\nAQB0QOgCAOiA0AUA0AGhCwCgA0IXAEAHhC4AgA4IXQAAHRC6AAA6IHQBAHRA6AIA6MBAoauqFlXV\nt6rq9qpa0mf/71bVuqq6qar+rqqOmrDvNVV1W+/rNcMsHgBgupizowOqau8klyd5UZKNSVZX1YrW\n2roJh/1Tkvmttfur6vVJ/ijJy6vqiUnelWR+kpZkbe/cHw/7jcBUMrZqbNQlADDFDNLTdWqS21tr\nd7bWHkxyVZJzJx7QWruutXZ/b/P6JIf3Xp+V5G9baz/qBa2/TbJoOKUDAEwfg4Suw5LcNWF7Y69t\ne347yWd28VwAgBlph8OLO6OqXpXxocTn7+R5i5MsTpIjjzxymCUBAEwJg4Suu5McMWH78F7bz6mq\nM5L8fpLnt9YemHDuwm3OXbXtua21ZUmWJcn8+fPbADUBMM30m+s4tnByG8xUgwwvrk5ydFXNq6rH\nJTk/yYqJB1TVyUk+mOSc1toPJuz6XJIzq+oJVfWEJGf22gAAZpUd9nS11rZU1YUZD0t7J7mitXZz\nVV2UZE1rbUWS9yQ5IMnHqypJvtNaO6e19qOqenfGg1uSXNRa+9EeeScAAFPYQHO6WmvXJrl2m7Z3\nTnh9xmOce0WSK3a1QACAmcCK9AAAHRjq3YvAuKUr14+6BACmGKELdtPcJddMbvQ3C4BtGF4EAOiA\n0AUA0AGhCwCgA0IXAEAHhC4AgA64xwp206Y5y0ddAgDTgNAFwMh4CDazieFFAIAOCF0AAB0QugAA\nOiB0AQB0QOgCAOiAuxcB6MTSlesntb3ljGNGUAmMhp4uAIAOCF0AAB0QugAAOmBOFwAj02+e19jC\n7uuALujpAgDogNAFANABoQsAoANCFwBAB4QuAIAOCF0AAB0QugAAOiB0AQB0wOKosBPmLrlmcqO/\nRQAMQE8XAEAHhC4AgA4YGAFgShlbNTa5beHkNphuhC7YCZvmLB91CQBMU4YXAQA6IHQBAHRA6AIA\n6IDQBQDQAaELAKAD7l6E7bD6PIzG0pXrJ7WNLey+Dhg2PV0AAB0QugAAOiB0AQB0QOgCAOiA0AUA\n0AH3YgEw5fW7m3jDpWePoBLYdXq6AAA6oKcLtmPTnOWjLgGAGURPFwBAB4QuAIAOCF0AAB0QugAA\nOiB0AQB0QOgCAOiA0AUA0IGB1umqqkVJ/iTJ3kn+rLV26Tb7T0+yNMmJSc5vrX1iwr6Hk3yjt/md\n1to5wygcgNmj37p5Y6tWT25bONZBNbBrdhi6qmrvJJcneVGSjUlWV9WK1tq6CYd9J8kFSd7W5xL/\n1lo7aQi1AgBMW4P0dJ2a5PbW2p1JUlVXJTk3ydbQ1Vrb0Nv3yB6oEQBg2hskdB2W5K4J2xuT/NpO\nfI/9qmpNki1JLm2t/c1OnAudGFs1NuoSAJjhunj24lGttbur6ilJ/r6qvtFau2PiAVW1OMniJDny\nyCM7KAkAoFuD3L14d5IjJmwf3msbSGvt7t6fdyZZleTkPscsa63Nb63NP/TQQwe9NADAtDFI6Fqd\n5OiqmldVj0tyfpIVg1y8qp5QVfv2Xj8pyXMyYS4YAMBsscPQ1VrbkuTCJJ9LckuSj7XWbq6qi6rq\nnCSpqgVVtTHJeUk+WFU3905/WpI1VfX1JNdlfE6X0AUAzDoDzelqrV2b5Npt2t454fXqjA87bnve\nV5OcsJs1wh63dOX6UZcAwAxnRXoAgA4IXQAAHRC6AAA6IHQBAHSgi8VRAWDo+t0AM7aw+zpgUHq6\nAAA6IHQBAHRA6AIA6IDQBQDQARPpmXXGVo2NugRgD5m75JpJbRsuPXsElcBkeroAADqgpwuAGWPT\nnOV9WvV0MTXo6QIA6IDQBQDQAaELAKADQhcAQAeELgCADghdAAAdELoAADogdAEAdMDiqMw6S1eu\nH3UJAMxCeroAADogdAEAdEDoAgDogDldAMxoc5dcM6ltw6Uegk339HQBAHRA6AIA6IDQBQDQAXO6\nAJjRNs1Z3qfVnC66p6cLAKADerqY0frdteSnHoBR0NMFANABoQsAoANCFwBAB4QuAIAOCF0AAB1w\nHxczWv/1eQCge3q6AAA6oKcLgFnn4He8clLbpouvHEElzCZ6ugAAOiB0AQB0QOgCAOiA0AUA0AGh\nCwCgA+5eZEaYu+Sa/jv8hAMwRfiVxIxgEVQApjrDiwAAHRC6AAA6YHiRaWds1dioSwCAnaanCwCg\nA3q6mHaWrlw/6hKAGajfXdAbLj17BJUwU+npAgDogNAFANABoQsAoAMDha6qWlRV36qq26tqSZ/9\np1fV16pqS1W9bJt9r6mq23pfrxlW4QAA08kOQ1dV7Z3k8iS/keS4JK+oquO2Oew7SS5IcuU25z4x\nybuS/FqSU5O8q6qesPtlAwBML4P0dJ2a5PbW2p2ttQeTXJXk3IkHtNY2tNZuSvLINueeleRvW2s/\naq39OMnfJlk0hLoBAKaVQULXYUnumrC9sdc2iIHOrarFVbWmqtbcc889A14aAGD6mBIT6Vtry1pr\n81tr8w899NBRlwMAMHSDhK67kxwxYfvwXtsgdudcAIAZY5DQtTrJ0VU1r6oel+T8JCsGvP7nkpxZ\nVU/oTaA/s9cGADCr7PAxQK21LVV1YcbD0t5Jrmit3VxVFyVZ01pbUVULkvx1kickeXFV/UFr7emt\ntR9V1bszHtyS5KLW2o/20HsBgF22ac7yPq0eA8TwVGtt1DX8nPnz57c1a9aMugymsIPf8cpRlwDM\nEgdv+a2+7Z7JyKOqam1rbf4gx06JifQAADOd0AUA0AGhCwCgA0IXAEAHdnj3IozS3CXXTG70UwvA\nNOTXF1Na/1u4AWD6EboAYDu2/z9+loxg55nTBQDQAaELAKADQhcAQAeELgCADghdAAAdELoAADog\ndAEAdMA6XUwZVp8HpouxVWOT2xZOboOJ9HQBAHRA6AIA6IDBG0aiX9d8sqDrMgCgM3q6AAA6oKeL\nKWP7D5YFgOlPTxcAQAf0dAHAEFhGgh0RugBgJy1duX5S21vOOGYElTCdGF4EAOiA0AUA0AHDiwAw\nBP2GHMcWdl8HU5eeLgCADujpYiT6/R8hAMxkeroAADogdAEAdMDwIntc/4dbA8DsoqcLAKADerrY\n40yaB2aruUuumdS24dKzR1AJU4GeLgCADghdAAAdELoAADogdAEAdEDoAgDogNAFANABoQsAoANC\nFwBAB4QuAIAOWJEeAPaQTXOW92m1Iv1spacLAKADeroYqn7PGfNTBgB6ugAAOqEPAgA6NLZqbHLb\nwsltzDyEcKGTAAANM0lEQVRCFwB0aOnK9ZPaxhZ2XwfdM7wIANABPV0MVf/bowEAPV0AAB0QugAA\nOmB4kV1mTS4AGJxfkQAwYpaRmB0GGl6sqkVV9a2qur2qlvTZv29VfbS3/x+ram6vfW5V/VtV3dj7\n+sBwywcAmB522NNVVXsnuTzJi5JsTLK6qla01tZNOOy3k/y4tfarVXV+kj9M8vLevjtaaycNuW4A\ngGllkJ6uU5Pc3lq7s7X2YJKrkpy7zTHnJvlw7/Unkrywqmp4ZQIATG+DhK7Dktw1YXtjr63vMa21\nLUl+kuSQ3r55VfVPVfWFqnpev29QVYurak1Vrbnnnnt26g0AAEwHe3oi/feSHNlau7eqnpnkb6rq\n6a21n048qLW2LMmyJJk/f37bwzUBwJTi0UCzwyA9XXcnOWLC9uG9tr7HVNWcJAclube19kBr7d4k\naa2tTXJHkmN2t2gAgOlmkJ6u1UmOrqp5GQ9X5yd55TbHrEjymiT/kORlSf6+tdaq6tAkP2qtPVxV\nT0lydJI7h1Y9I+WRPwB7jmUkZp4dhq7W2paqujDJ55LsneSK1trNVXVRkjWttRVJPpTkL6rq9iQ/\nyngwS5LTk1xUVQ8leSTJ61prP9oTb4Q96+B3bJuzAYCdMdCcrtbatUmu3abtnRNeb05yXp/zrk5y\n9W7WCAAw7VmRHgCmIJPrZx6hi0n6zSMAAHbPQI8BAgBg9whdAAAdMLzIJP3mEQAAu0dPFwBAB/R0\nzXJzl1wzudFPBcCU1O/f7A2Xnj2CStgVfr0CwDTR/0kgQtd0YXgRAKADerpmOc9PBIBu6OkCAOiA\n0AUA0AGhCwCgA0IXAEAHTKQHgGlsbNXY5LaFk9sYPaELAKax/o9uG5vcIoiNnNA1i1h9HgBGx5wu\nAIAO6OeYRSyECgCjo6cLAKADQhcAQAcMLwLADOOOxqlJTxcAQAf0dM1Q/RbLAwBGR+gCgFmg35Dj\n2MLu65jNhK4ZoN+ipxcsGkEhAMB2CV0zVP9JlADAqAhdADBLeVh2t4SuGcBK8wAw9Qld00jfB1Yn\n/isCMDR6v/Ycv64BYJbqN//3LWccM4JKZgehCwDYatDV7BM9YDvLivQAAB0QugAAOmB4cYrqN5Fx\n0xxrbwHAdCV0AQCPaXsLbnuM0M4RuqaAfktB6NUCgJlF6AIAdsnB73jlpLZNF185gkqmB6FrCrCi\nPADMfEIXADA0VrTfPqFrgi5+UPp1xQLATDHo4qqzMYgJXRP0+0FxZwYAMAwWRwUA6ICerj3IUCIA\nGHJ8lNAFAEwJM30SvtC1A/0WLt1w6dkDHefTBQAeJRbsgv4ryFtrCwAGtb1HC21rJt3QZiI9AEAH\n9HTtgB4sABidmTTPS+gCAKasmXTno9AFAEwr03Uxc6ELAJj2psMwpNAFAEx706H3a6C7F6tqUVV9\nq6pur6olffbvW1Uf7e3/x6qaO2Hf7/Xav1VVZw2vdACA6WOHPV1VtXeSy5O8KMnGJKurakVrbd2E\nw347yY9ba79aVecn+cMkL6+q45Kcn+TpSZ6cZGVVHdNae3jYb2RneUQPANClQXq6Tk1ye2vtztba\ng0muSnLuNsecm+TDvdefSPLCqqpe+1WttQdaa99OcnvvegAAs8ogoeuwJHdN2N7Ya+t7TGttS5Kf\nJDlkwHMBAGa8KTGRvqoWJ1nc27yvqr7Vwbd9UpIfdvB9Zguf5/D5TIfL5zl8PtPh85kOUV3ykS4+\nz6MGPXCQ0HV3kiMmbB/ea+t3zMaqmpPkoCT3DnhuWmvLkiwbtOhhqKo1rbX5XX7PmcznOXw+0+Hy\neQ6fz3T4fKbDNdU+z0GGF1cnObqq5lXV4zI+MX7FNsesSPKa3uuXJfn71lrrtZ/fu7txXpKjk9ww\nnNIBAKaPHfZ0tda2VNWFST6XZO8kV7TWbq6qi5Ksaa2tSPKhJH9RVbcn+VHGg1l6x30sybokW5K8\ncSrcuQgA0LWB5nS11q5Ncu02be+c8HpzkvO2c+4lSS7ZjRr3lE6HM2cBn+fw+UyHy+c5fD7T4fOZ\nDteU+jxrfBQQAIA9aaAV6QEA2D2zOnRV1bur6qaqurGqPl9VTx51TdNZVb2nqm7tfaZ/XVUHj7qm\n6a6qzquqm6vqkaqaMnfgTDc7epQZO6eqrqiqH1TVN0ddy0xQVUdU1XVVta739/3No65puquq/arq\nhqr6eu8z/YNR15TM8uHFqvqF1tpPe6/flOS41trrRlzWtFVVZ2b8ztUtVfWHSdJa+28jLmtaq6qn\nJXkkyQeTvK21tmbEJU07vUeZrc+ER5klecU2jzJjJ1TV6UnuS/L/tdaOH3U9011V/XKSX26tfa2q\nDkyyNslL/Izuut5TcfZvrd1XVfsk+XKSN7fWrh9lXbO6p+vRwNWzf5LZm0CHoLX2+d4TCZLk+oyv\ny8ZuaK3d0lrrYrHgmWyQR5mxE1prX8z4neoMQWvte621r/Ve/yzJLfH0lt3Sxt3X29yn9zXy3/Gz\nOnQlSVVdUlV3JfmtJO/c0fEM7D8n+cyoi4B4HBnTSFXNTXJykn8cbSXTX1XtXVU3JvlBkr9trY38\nM53xoauqVlbVN/t8nZskrbXfb60dkWR5kgtHW+3Ut6PPs3fM72d8Xbblo6t0+hjkMwVmvqo6IMnV\nSd6yzUgMu6C19nBr7aSMj7qcWlUjHwqfEs9e3JNaa2cMeOjyjK9F9q49WM60t6PPs6ouSPIfk7yw\nzeYJgzthJ35G2TUDPY4MRqk37+jqJMtba3816npmktbapqq6LsmiJCO9+WPG93Q9lqo6esLmuUlu\nHVUtM0FVLUryX5Oc01q7f9T1QM8gjzKDkelN+v5Qkltaa+8ddT0zQVUd+ugd9FX1HzJ+I83If8fP\n9rsXr07y1IzfHfbPSV7XWvN/wLuo9xiofTP+sPMkud7doLunql6a5P1JDk2yKcmNrbWzRlvV9FNV\n/2uSpfn3R5lNxadkTBtV9ZEkC5M8Kcn3k7yrtfahkRY1jVXVc5N8Kck3Mv77KEne3nsaDLugqk5M\n8uGM/53fK8nHWmsXjbaqWR66AAC6MquHFwEAuiJ0AQB0QOgCAOiA0AUA0AGhCwCgA0IXsEuq6n+p\nqquq6o6qWltV11bVMXvw+311T127z/e6b1ePqarXVdWre6//vKpetoPr9D2mqp5cVZ8YtGZg6pvx\nK9IDw9dbzPGvk3y4tXZ+r+0ZSX4pyfohf685rbUtrbVnD/O6e0pr7QNDus53kzxmYAOmFz1dwK54\nQZKHJgaM1trXk3y5qt7Te3bkN6rq5UnS6xE7+9FjH+3dqaq5VfWlqvpa7+vZvf0Le+0rkqzrtd3X\n+/OAqvq73vHfePQZlb1r3VJV/7Oqbq6qz/dWok5V/WrvGZdf7533K732/1JVq6vqpqr6g23fZK+O\nL1bVNVX1rar6QFXtNWH/Jb1rXl9Vv9RrG6uqt/W51oaq+qNezTdU1a9O2H16VX21qu58tNer935G\n+sgSYLiELmBXHJ9kbZ/2/y3JSUmekeSMJO+pql9O8tEk/3uS9B7F88Ik1yT5QZIXtdZOSfLyJO+b\ncK1Tkry5tbbtkOXmJC/tnfOCJP9Pr+ctSY5Ocnlr7ekZX8H/N3vty3vtz0jy7CTfq6oze8ef2qv5\nmVV1ep/3dGqS/zPJcUl+pfcek2T/jD914RlJvpjktdv5rCb6SWvthCT/b8ZXyH/ULyd5bsafW3rp\nANcBpiGhCxim5yb5SGvt4dba95N8IcmCJJ9J8oKq2jfJbyT5Ymvt35Lsk+R/VtU3knw848HmUTe0\n1r7d53tUkv9eVTclWZnksIwPaybJt1trN/Zer00yt6oOTHJYa+2vk6S1trn3bNAze1//lORrSY7N\neAjb1g2ttTtbaw8n+UjvPSbJg0k+PfF7DfD5fGTCn8+a0P43rbVHWmvrJrwXYIYxpwvYFTdnJ+Yb\ntdY2V9WqJGdlvEfrqt6ut2b82X3PyPj/BG6ecNq/budyv5XxZ1E+s7X2UFVtSLJfb98DE457OMl/\neIyyKsn/3Vr74I7K3872Q+3fn6P2cAb797Rt5/XEuivAjKSnC9gVf59k36pa/GhD7wGzm5K8vKr2\nrqpDk5ye5IbeIR9N8n8keV6Sz/baDkryvdbaI0n+U8YfTrsjByX5QS9wvSDJUY91cGvtZ0k2VtVL\nenXuW1WPT/K5JP+5qg7otR9WVb/Y5xKnVtW83lyulyf58gA1bs/LJ/z5D7txHWAa0tMF7LTWWquq\nlyZZWlX/LeM9VBuSvCXJAUm+nvGenP/aWvuX3mmfT/IXST7ZWnuw1/anSa7uLbHw2Wy/d2ui5Uk+\n1RuSXJPk1gHO+U9JPlhVFyV5KMl5rbXPV9XTkvxDb0rYfUlelfF5ZhOtzvgcrF9Ncl3G79rcVU/o\nDYs+kOQVu3EdYBqqf+8dB2CiqlqY5G2ttf84hGttSDK/tfbD3b0WMD0ZXgQA6ICeLgCADujpAgDo\ngNAFANABoQsAoANCFwBAB4QuAIAOCF0AAB34/wFs6lFsD437XgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b2bf6cec358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for ic in range(len(c_names)):\n",
    "    plotting.plot_hists(data_c[:,0,ic],mc_c[:,0,ic],generated_w=mc_w,bins=100,range=[-2.0,2.0])\n",
    "    plt.xlabel(c_names[ic])\n",
    "    plt.show()\n",
    "\n",
    "for ix in range(len(x_names)):\n",
    "    plotting.plot_hists(data_x[:,0,ix],mc_x[:,0,ix],generated_w=mc_w,bins=100,range=[-3,3])\n",
    "    plt.xlabel(x_names[ix])\n",
    "    plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prepare train and test sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "nmax = min(data_x.shape[0]//FRAC_DATA,mc_x.shape[0])\n",
    "\n",
    "data_x_train,data_x_test,data_c_train,data_c_test,data_w_train,data_w_test = cms.train_test_split(data_x[:nmax],data_c[:nmax],data_w[:nmax],test_size=0.1)\n",
    "mc_x_train,mc_x_test,mc_c_train,mc_c_test,mc_w_train,mc_w_test = cms.train_test_split(mc_x[:nmax],mc_c[:nmax],mc_w[:nmax],test_size=0.1)\n",
    "\n",
    "wscl = data_w[:nmax].sum() / mc_w[:nmax].sum()\n",
    "mc_w_train *= wscl\n",
    "mc_w_test *= wscl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(nmax,mc_w_train.sum()/data_w_train.sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Instantiate the model\n",
    "\n",
    "Create the model and compile it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reload(models)\n",
    "\n",
    "xz_shape = (1,len(x_names))\n",
    "c_shape = (1,len(c_names))\n",
    "gan = models.MyFFGAN( xz_shape, xz_shape, c_shape=c_shape,\n",
    "                     g_opts=G_OPTS,\n",
    "                     d_opts=D_OPTS,\n",
    "                     dm_opts=DM_OPTS,\n",
    "                     am_opts=AM_OPTS,\n",
    "                     gan_targets=GAN_TARGETS\n",
    "                    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This will actually trigger the instantiation of the generator\n",
    "(if not done here, it will happen befor compilation)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 7)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.engine.training.Model at 0x2b2c3ec2f240>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gan.get_generator()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Same as above for the discriminator.  \n",
    "Two discriminator models are created: one for the discriminator training phase and another\n",
    "for the generator trainin phase.  \n",
    "The difference between the two is that the second does not contain dropout layers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WeightClip 0.02\n",
      "WeightClip 0.02\n",
      "WeightClip 0.02\n",
      "WeightClip 0.02\n",
      "WeightClip 0.02\n",
      "WeightClip 0.02\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(<keras.engine.training.Model at 0x2b2c3eecba20>,\n",
       " <keras.engine.training.Model at 0x2b2c3eecbc18>)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gan.get_discriminator()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### model compilation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "wgan_loss\n"
     ]
    }
   ],
   "source": [
    "gan.adversarial_compile(loss=LOSS,schedule=SCHEDULE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "G_512to64x8_c_input (InputLayer) (None, 1, 5)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_input (InputLayer)   (None, 1, 7)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_all_inputs (Concaten (None, 1, 12)         0           G_512to64x8_c_input[0][0]        \n",
      "                                                                   G_512to64x8_input[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up1_dense (Dense)    (None, 1, 512)        6656        G_512to64x8_all_inputs[0][0]     \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up1_activ (PReLU)    (None, 1, 512)        512         G_512to64x8_up1_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up2_dense (Dense)    (None, 1, 512)        262656      G_512to64x8_up1_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up2_activ (PReLU)    (None, 1, 512)        512         G_512to64x8_up2_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up3_dense (Dense)    (None, 1, 256)        131328      G_512to64x8_up2_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up3_activ (PReLU)    (None, 1, 256)        256         G_512to64x8_up3_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up4_dense (Dense)    (None, 1, 256)        65792       G_512to64x8_up3_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up4_activ (PReLU)    (None, 1, 256)        256         G_512to64x8_up4_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up5_dense (Dense)    (None, 1, 128)        32896       G_512to64x8_up4_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up5_activ (PReLU)    (None, 1, 128)        128         G_512to64x8_up5_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up6_dense (Dense)    (None, 1, 128)        16512       G_512to64x8_up5_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up6_activ (PReLU)    (None, 1, 128)        128         G_512to64x8_up6_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up7_dense (Dense)    (None, 1, 64)         8256        G_512to64x8_up6_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up7_activ (PReLU)    (None, 1, 64)         64          G_512to64x8_up7_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up8_dense (Dense)    (None, 1, 64)         4160        G_512to64x8_up7_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_up8_activ (PReLU)    (None, 1, 64)         64          G_512to64x8_up8_dense[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_output (Dense)       (None, 1, 7)          455         G_512to64x8_up8_activ[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "G_512to64x8_add (Add)            (None, 1, 7)          0           G_512to64x8_input[0][0]          \n",
      "                                                                   G_512to64x8_output[0][0]         \n",
      "====================================================================================================\n",
      "Total params: 530,631\n",
      "Trainable params: 530,631\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "gan.get_generator().summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "D_1024to32x6_c_input (InputLayer (None, 1, 5)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_input (InputLayer)  (None, 1, 7)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_all_inputs (Concate (None, 1, 12)         0           D_1024to32x6_c_input[0][0]       \n",
      "                                                                   D_1024to32x6_input[0][0]         \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down1_dense (Dense) (None, 1, 1024)       13312       D_1024to32x6_all_inputs[0][0]    \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down1_activ (Activa (None, 1, 1024)       0           D_1024to32x6_down1_dense[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down2_dense (Dense) (None, 1, 512)        524800      D_1024to32x6_down1_activ[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down2_activ (Activa (None, 1, 512)        0           D_1024to32x6_down2_dense[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down3_dense (Dense) (None, 1, 128)        65664       D_1024to32x6_down2_activ[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down3_activ (Activa (None, 1, 128)        0           D_1024to32x6_down3_dense[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down4_dense (Dense) (None, 1, 64)         8256        D_1024to32x6_down3_activ[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down4_activ (Activa (None, 1, 64)         0           D_1024to32x6_down4_dense[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down5_dense (Dense) (None, 1, 32)         2080        D_1024to32x6_down4_activ[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down5_activ (Activa (None, 1, 32)         0           D_1024to32x6_down5_dense[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_flat (Flatten)      (None, 32)            0           D_1024to32x6_down5_activ[0][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_output (Dense)      (None, 1)             33          D_1024to32x6_flat[0][0]          \n",
      "====================================================================================================\n",
      "Total params: 614,145\n",
      "Trainable params: 614,145\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "gan.get_discriminator()[0].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "D_1024to32x6_c_input (InputLayer (None, 1, 5)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_input (InputLayer)  (None, 1, 7)          0                                            \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_all_inputs (Concate (None, 1, 12)         0           D_1024to32x6_c_input[0][0]       \n",
      "                                                                   D_1024to32x6_input[0][0]         \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down1_dense (Dense) (None, 1, 1024)       13312       D_1024to32x6_all_inputs[0][0]    \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down1_activ (Activa (None, 1, 1024)       0           D_1024to32x6_down1_dense[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down2_dense (Dense) (None, 1, 512)        524800      D_1024to32x6_down1_activ[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down2_activ (Activa (None, 1, 512)        0           D_1024to32x6_down2_dense[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down3_dense (Dense) (None, 1, 128)        65664       D_1024to32x6_down2_activ[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down3_activ (Activa (None, 1, 128)        0           D_1024to32x6_down3_dense[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down4_dense (Dense) (None, 1, 64)         8256        D_1024to32x6_down3_activ[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down4_activ (Activa (None, 1, 64)         0           D_1024to32x6_down4_dense[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down5_dense (Dense) (None, 1, 32)         2080        D_1024to32x6_down4_activ[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_down5_activ (Activa (None, 1, 32)         0           D_1024to32x6_down5_dense[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_flat (Flatten)      (None, 32)            0           D_1024to32x6_down5_activ[1][0]   \n",
      "____________________________________________________________________________________________________\n",
      "D_1024to32x6_output (Dense)      (None, 1)             33          D_1024to32x6_flat[1][0]          \n",
      "====================================================================================================\n",
      "Total params: 614,145\n",
      "Trainable params: 614,145\n",
      "Non-trainable params: 0\n",
      "____________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "gan.get_discriminator()[1].summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "set()"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# sanity check\n",
    "set(gan.am[0].trainable_weights)-set(gan.am[1].trainable_weights)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training\n",
    "\n",
    "Everything is ready. We start training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.optimizers import RMSprop\n",
    "\n",
    "if PRETRAIN_G:\n",
    "    generator = gan.get_generator()\n",
    "    generator.compile(loss=\"mse\",optimizer=RMSprop(lr=1.e-3))\n",
    "    generator.fit( [mc_c_train,mc_x_train], [mc_c_train,mc_x_train], sample_weight=[mc_w_train,mc_w_train], epochs=1, batch_size=BATCH_SIZE  )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# reload(base)\n",
    "\n",
    "initial_epoch = 0\n",
    "# if hasattr(gan.model,\"history\"):\n",
    "#     initial_epoch = gan.model.history.epoch[-1] + 1\n",
    "\n",
    "do = dict(\n",
    "    x_train=data_x_train,\n",
    "    z_train=mc_x_train,\n",
    "    c_x_train=data_c_train,\n",
    "    c_z_train=mc_c_train,\n",
    "    w_x_train = data_w_train,\n",
    "    w_z_train = mc_w_train,\n",
    "          \n",
    "    x_test=data_x_test,\n",
    "    z_test=mc_x_test,\n",
    "    c_x_test=data_c_test,\n",
    "    c_z_test=mc_c_test,\n",
    "    w_x_test = data_w_test,\n",
    "    w_z_test = mc_w_test,\n",
    "    \n",
    "    n_epochs=EPOCHS + initial_epoch - 1,\n",
    "    initial_epoch=initial_epoch,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    plot_every=PLOT_EVERY,\n",
    "    \n",
    "    monitor_dir=MONITOR_DIR\n",
    ")\n",
    "\n",
    "exc = None\n",
    "try:\n",
    "    gan.adversarial_fit(**do)\n",
    "except Exception as e:\n",
    "    exc = e"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not exc is None:\n",
    "    print(exc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}