physbo_try.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0bc27be4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from rdkit import rdBase, Chem\n",
    "from rdkit.Chem import AllChem, Draw, PandasTools, Descriptors\n",
    "from rdkit.Chem.Draw import IPythonConsole\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "sns.set()\n",
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "import statsmodels.formula.api as smf\n",
    "import statsmodels\n",
    "import urllib.request \n",
    "url = 'https://raw.githubusercontent.com/HIPS/neural-fingerprint/master/data/2015-05-24-delaney/ci034243xsi20040112_053635.txt'\n",
    "urllib.request.urlretrieve(url, 'water_solubility.txt') \n",
    "df = pd.read_csv('water_solubility.txt', sep=',') \n",
    "PandasTools.AddMoleculeColumnToFrame(frame=df, smilesCol='smiles')\n",
    "df.columns = ['ID', 'm_sol', 'p_sol', 'SMILES', 'ROMol']\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d7b4337b",
   "metadata": {},
   "outputs": [],
   "source": [
    "mols = [Chem.MolFromSmiles(i) for i in df['SMILES']]\n",
    "\n",
    "maccskeys = []\n",
    "for m in mols:\n",
    "    maccskey = [x for x in AllChem.GetMACCSKeysFingerprint(m)]\n",
    "    maccskeys.append(maccskey)\n",
    "X = np.array(maccskeys)\n",
    "y = df['m_sol']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "38fa7bcb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.3694205637413728"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "regr = RandomForestRegressor(n_estimators=100,random_state=0)\n",
    "regr.fit(X, y)\n",
    "mean_absolute_error(regr.predict(X),y)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bf2c81b2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "b9330e9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def logp_calc(x):\n",
    "    return regr.predict(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "cbeeb2fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "from numpy.random import randint\n",
    "from numpy.random import rand\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "592c3fab",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "9ec19cfe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9749296772323556"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rand()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "d935cd10",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(233)\n",
    "def mutation(pop,r_mut):\n",
    "    for i in range(len(pop)):\n",
    "        if rand() > r_mut:\n",
    "            pop[i] = randint(0,2)\n",
    "    return pop\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "6bcbe09c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 167)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xcan = []\n",
    "for i in range(10000):\n",
    "    np.random.seed(i)\n",
    "    xcan.append([random.randint(0, 1) for p in range(167)])\n",
    "xcan = np.array(xcan)\n",
    "xcan.shape    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "dc2aa1e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "815107aa",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "2b9d8b7d",
   "metadata": {},
   "outputs": [],
   "source": [
    "class simulator(object):\n",
    "    def __init__(self, x):\n",
    "        self.t = -logp_calc(x)\n",
    "\n",
    "    def __call__( self, action):\n",
    "        return self.t[action]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a5cf0f9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "id": "efc073cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "import physbo\n",
    "simu = simulator(xcan)\n",
    "\n",
    "policy = physbo.search.discrete.policy(test_X=xcan)\n",
    "policy.set_seed(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "id": "1115d127",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0001-th step: f(x) = 2.307077 (action=9394)\n",
      "   current best f(x) = 2.307077 (best action=9394) \n",
      "\n",
      "0002-th step: f(x) = 4.045877 (action=3583)\n",
      "   current best f(x) = 4.045877 (best action=3583) \n",
      "\n",
      "0003-th step: f(x) = 4.644623 (action=4358)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n",
      "0004-th step: f(x) = 3.634893 (action=2991)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n",
      "0005-th step: f(x) = 3.951923 (action=2323)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n"
     ]
    }
   ],
   "source": [
    "res = policy.random_search(max_num_probes=5, simulator=simulator(xcan))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2e491226",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "id": "992e8398",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start the initial hyper parameter searching ...\n",
      "Done\n",
      "\n",
      "Start the hyper parameter learning ...\n",
      "0 -th epoch marginal likelihood 5.973744255704563\n",
      "50 -th epoch marginal likelihood 5.911119598787423\n",
      "100 -th epoch marginal likelihood 5.871459819719742\n",
      "150 -th epoch marginal likelihood 5.849656932241771\n",
      "200 -th epoch marginal likelihood 5.839271121890324\n",
      "250 -th epoch marginal likelihood 5.835026566315912\n",
      "300 -th epoch marginal likelihood 5.833545510633604\n",
      "350 -th epoch marginal likelihood 5.8331046241982785\n",
      "400 -th epoch marginal likelihood 5.832992540391977\n",
      "450 -th epoch marginal likelihood 5.832968185802832\n",
      "500 -th epoch marginal likelihood 5.832963664575456\n",
      "Done\n",
      "\n",
      "0006-th step: f(x) = 2.975433 (action=2265)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n",
      "0007-th step: f(x) = 3.807606 (action=5403)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n",
      "0008-th step: f(x) = 3.615812 (action=4499)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n",
      "0009-th step: f(x) = 4.328649 (action=4227)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n",
      "0010-th step: f(x) = 3.454812 (action=4677)\n",
      "   current best f(x) = 4.644623 (best action=4358) \n",
      "\n"
     ]
    }
   ],
   "source": [
    "res_TS = policy.bayes_search(max_num_probes=5, \n",
    "                             simulator=simulator(xcan), score='TS',\n",
    "                             interval=10, \n",
    "                             num_rand_basis=5000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9f8a6d9a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "d9550a9d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([9394, 3583, 4358, 2991, 2323, 4050, 3708,  505, 5163, 9198, 1304,\n",
       "       2988, 8424, 5710, 3382, 1634, 7774, 3007, 3087,  869, 8709, 1414,\n",
       "       9267, 1811, 3006, 2743, 4555, 5780, 9022, 8628, 2115, 9388,  387,\n",
       "       3516, 4440, 3601, 2380, 7514,  876,  918])"
      ]
     },
     "execution_count": 202,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_TS.chosen_actions[:40]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "id": "96830724",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-3.40654908])"
      ]
     },
     "execution_count": 206,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_calc(xcan[918].reshape(1, -1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "b9ea629f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([-2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667]),\n",
       " array([9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394,\n",
       "        9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394,\n",
       "        9394, 9394, 9394]))"
      ]
     },
     "execution_count": 174,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_TS.export_sequence_best_fx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "f4a0f6dc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([-2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667]),\n",
       " array([9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394,\n",
       "        9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394,\n",
       "        9394, 9394, 9394]))"
      ]
     },
     "execution_count": 175,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_TS.export_all_sequence_best_fx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "e4f620f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "4195f040",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res.total_num_search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "ebee67a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f9b2e4d1b20>]"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAD7CAYAAABpJS8eAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA9YUlEQVR4nO3deXhb13ng/+8FQAIEuAPgAoqkJMo6WqjFlmRrseUkXpLYSbokTppkErfZp/t02s6vTTvTZJ5p59cmkzxNl3SSJmnrpkmdTjJNbMdOYieWbEuWF+3SkUVSlERw30CCO4H5AwAFUSAJYgfxfp7Hj8WLC9yjS+rlue855z1GMBhECCHE2mfKdgOEEEJkhgR8IYQoEBLwhRCiQEjAF0KIAiEBXwghCoQl2w1YghXYB3QD81luixBC5AszUA+cAKYXv5irAX8fcCTbjRBCiDx1D3B08cFcDfjdAMPDfgKBxNYJOJ2lDA6Op7RRqSTtS460LznSvuTkavtMJoOqKgeEY+hiuRrw5wECgWDCAT/y/lwm7UuOtC850r7k5Hj7YqbCZdBWCCEKhAR8IYQoEBLwhRCiQEjAF0KIAiEBXwghCoQEfCGEKBBrLuDPzQf49FeOcfJSX7abIoQQOWXNBXzDgL7hSU5fHsh2U4QQIqesuYBvNpmoKrPSOzSR7aYIIUROWXMBH8BdWUKfBHwhhLjJmgz4zgqb9PCFEGKRNRnwXRU2hsemmZmVyspCCBGxJgO+u6IEgEHfVJZbIoQQuWNNBnxnhQ2AgVEJ+EIIEbEmA767MtTDl4AvhBA3rMmAX1FajMVsYmBkMttNEUKInLEmA77JMKipKpEevhBCREl4xyul1CHgC0AxMAh8RGvdueicZuAs0BY+1Ku1fmui11yNmmo7A6PSwxdCiIhktjj8Z+BdWuvTSqmPAH8J/Nyic/YC39RafzKJ6ySkttpO2/WRTF9WCCFyVkIpHaWUFfgjrfXp8KHTQFOMU/cBrUqpk0qpZ5VSOxJs56rVVtsZm5hlamYuU5cUQoicllDA11pPa60fA1BKmYA/Ab4X49Qp4DHgDuBzwPeUUsUJtXSVaqrsAAxKHl8IIYA4UjpKqUcI5eqjXdRa3x8O3v8Q/pw/XfxerfWfRH35pFLqz4CtwKl4Gud0lsZzWky1E7MAzGDgdpcl/DnplKvtipD2JUfalxxpX+qtGPC11o8Djy8+rpQqBf6d0IDtz2mtZ2Oc8xuEcviD4UMGcMt5SxkcHCcQCMZ7+k1qwz389qvDbHA7EvqMdHK7y+jvH8t2M5Yk7UuOtC850r7EmEzGsh3lZKZlPgZcBt6ntZ5e4px7gY8CKKXuBczAxSSuGbfKMitFFpPM1BFCiLCEZukopW4nNCPnPPCaUgrAq7V+SCn1KcCjtf6vwG8B31BKfRiYBN6vtQ6kpunLMwwDV4WNgRHJ4QshBCQY8LXWrxNKz8R67ctRf+4CHkisaclzVcjiKyGEiFiTK20jXBU2SekIIUTYmg/4/qk5JqZkLr4QQqztgL9QNVN6+UIIsbYDfrguviy+EkKIAgn4/RLwhRBibQf80pIirEVmSekIIQRrPOAbhoGr0iYpHSGEYI0HfABXuY1+WXwlhBAFEPArSxj0TRIMJlaTRwgh1oq1H/ArbExOz+OXufhCiAJXEAEfZGqmEEIUQMAPLb7qH5GZOkKIwrb2A35lqIcvRdSEEIVuzQd8h62IEqtF5uILIQremg/4EKmaKT18IURhK5iAL4O2QohCVyABv4T+UZmLL4QobAUS8G3MzAYYm4x7/3QhCtaTxzp56WxPtpsh0iChLQ7zzcJMnZEpyu3FWW6NELlrPhDg31/owFZsYd/WGizmgugTFoyC+G5G5uLLTB0hlne9z8/MbACff4YzbYPZbo5IsQIJ+DIXX4h4tHtHAbAVm3n+lDfLrRGpVhABv8RqwWGzSMAXYgWXu3yUO4p5yx3rON0+yPDYdLabJFKoIAI+hKpmDkh5BSGW1eYdpcVTzj076wkG4YUz3dlukkihwgn4svhKiGWNTczQNzxJS0MFtdV2VGMlR093E5DpzGtGwQR8d0UJg74pmYsvxBLavD4AWjzlABze5aFvZBJ9dSSLrRKpVDAB31lhY3YuwKh/JttNESIntXWNYjYZrK8PBfw9yk2J1cKR0zJ4u1YUTMCXmTpCLK+ta5R1NaVYi8wAFBeZ2b+9llcu9uOfkkWLa0HCC6+UUvcAXwSKgQ7gUa318KJzioG/B/YCk8AHtNYXE25tElyV4bn4I5NsaqjIRhOEyFmBQJCO7jEO7ai76fjhnR6ee62LY+d6uW/Puiy1TqRKMj38rwMf0lrvAM4DvxfjnN8E/FrrrcBvA99I4npJcZVLD1+IpVzvH2d6dp6WRZ2h5roymmpLef6UV8a/1oBkAv5WrfV5pVQR0AAMxzjnYeCfAbTWzwNupVRTEtdMmLXYTLm9SFbbChHDwoBtjKffw7s8XOsbp7N3LNPNEimWcMDXWs8qpXYA14E3A9+KcZoHiJ7I2w1k7bnQWVEiPXwhYmjvGqXMXoQ7PNYVbf+2WoosJo6ckjn5+W7FHL5S6hHgC4sOX9Ra36+1PgPUKqU+CXwbOLjoPBMQ/RxoAIF4G+d0lsZ7akxud9lNX6+rLePytZFbjmdLrrRjKdK+5ORT+670jrFtg5OamvKY5x7a5eH4uR5+9b27sRVnpuZiPt2/fLHid05r/TjwePQxpZRNKfXzWuvvhQ89Bnw+xtuvA/VAW/jrOiDuOV6Dg+MEAonlDd3uMvr7b34ELbWZ6RueoLfXh8lkJPS5qRKrfblE2pecfGrf+OQsXf1+9m+rXbLNdyk3P331Ok+/0M7B1vqMti8X5Wr7TCZj2Y5yoimdWeCvlVJ7wl+/Fzga47wngQ8DKKXuBqa01lcTvGbSXBUlzAeCjIxLfRAhItq6QgXTlpu9trmxkpqqEp6XtE5eSyjga63ngfcB/1spdRJ4D/AxAKXUp5RSnw2f+iXAqpQ6B/wl8KGkW5wEt8zFF+IWbV4fJsNgfV3sdA6AYRjcs7OeS9dG6BmayGDrRColnIzTWh8F9sQ4/uWoP08BjyZ6jVRzLgT8STY3Vma3MULkiNCCKwfWYvOy5x3aUc93n+/gyGkvj7xpU4ZaJ1KpYFbaQtRq2xHp4QsBoQVX7d2+mNMxF6sstbKzxckLZ3qYm4977oXIIQUV8IssZipKiyWlI0SYd8DP9Mz8QsG0ldyzq152w8pjBRXwIVQ1UxZfCRFyObzDVTw9fICdLU4qSotlN6w8VXABX+riC3FDW9copSVF1IRrTa3EbDJx94562Q0rTxVewK+0MeSbZj4gOUgh2r0+WjzlGEb861Lult2w8lbhBfyKEgLBIMM+6Z2IwjY+OUv34ETc6ZyI2io7W5oqOXLaK7th5ZmCC/hOmYsvBBDq3UP8+fto9+z00D8yJbth5ZmCC/iRxVf9MnArClxb1yiGARvqV18TZmE3LBm8zSsFF/Cry20YwKD08EWBa/eOss5dmlAxtIXdsLTshpVPCi7gW8wmqsqt9MviK1HAVrPgaimHd3qYmw9w7FxvClsm0qngAj6Edr8alJROTIFgkM9/63X+/388waVrI7LL0Rp1rXeMyen4F1zF0lxXRnNtmeyGlUcyU9g6x7gqS7jQGWuDLtHh9XHuyjAW8whHT3lpqinlvj3ruGtbLcVFy9daEfnjYvjnP5kePoRW3j72zCU6e8eWLb4mckNh9vArbIyMTUs9kBhOXh7AZBh89dMP8OjbFPPBIF9/6iK/+zcv8p2ftsnYxxqhO4coLSmitiq+BVdLieyGJWWT80NB9vCdFTaCwJBvipoqe7abk1NOXR7ktnUVOCtKuHd3A4d3edBXR/jJq9d56ngnTx3v5I7b3Ny3Zx2qqXJVC3ZE7rjYOcTGVS64isVuK2KvcnP8fA/ve8smrPIUmNMKMuC7K0K9mv5RCfjRBkYnud4/znvffKP0rWEYbGmuYktzFQOjkzz3ehfPn/Ty6qV+1rkd3LdnHfu318k/9DwyMTXLtd5x9qqalHze4V0eXjrXyysX+zi0I/27YYnEFWxKB2Rq5mKnLocqIO7a5Iz5uquihEfetInP/9ohfuXtWzAZBv/wQ83v/vUL/OuzlxkYkYHwfLCw4CqJAdtokd2wjpyWtE6uK8geflW5FZNh0C8B6ian2gaorSqh3ulY9rziIjP37PJw98563rg+yo9fvc4zJ67x9Imr7N7k4r4969jaXCXpnhx1uWsUkwEb6lMT8CO7Yf3bz9rpGZqgrlqemnNVQfbwzSYT1eVW6eFHmZqZ42LnMLs2ueJ+j2EYbG6s5Fd/vpU//48HePhAM5e7Rvnct07y1PGsbV0sVtDm9dFUV06JNXX9vUM76jEZBkdOy8rbXFaQAR9CaR0pr3DDuY5h5uaDqwr40arLbfzi4RY+96sH2dJUyXOvdUlhrRwUCAZp9/pQzVUp/VzZDSs/FHDAL5ECalFOXR6gxGrhtnXJzcsuspg5vMvDoG+KN66NpKZxq/Svz13mTLvsyBRL9+AEk9NzbGmuTvlnH97lweef4bTshpWzCjfgV9oYHZ9hZnY+203JukAwyOm2AXZsrMZiTv5H4vbNbqzFZl4825OC1q3O1d4xfnj8qgwgLqGtK7TD1Zb1qe3hA+xoqaaitFgKquWwwg34kZk6Punld3T78E3MJpzOWcxaZGavcvOK7sv4L9Sj4UDfPeDP6HXzRbt3FIfNgsdVmvLPlt2wcl8BB/zQXHwZuA2lc0yGwY6NsadjJuLg9jomp+c5eXkgZZ+5ktm5AC+dCz1V9AxNyK5mMbR1+djoqcBkSs8MqshuWEdlN6ycVMABP1IXXwL+qcuDbFpXQWlJUco+UzVXUVVmzWha5/U3+vFPzbF/Wy3zgSB9wzIoH21iag7vgJ+WhvTVvKmtsrOhvpxzMoaSkwo24FeWWjGbDAYKfKbO4OgU1/rG2Z2idE6EyTA4sL2Os+1D+PwzKf3spRw53Y2z3Mp9e9YBoQFKcUN79yhBki+YtpL1dWVc6/dLBc0cVLAB32QycFbYGCjwuvin2kIpl6VW1ybjQGsdgWCQ4+fTXy99cHSK8x1DHNpRj8cVWjjmlTz+Tdq7fBjAxhQtuFpKY20pk9Nzki7NQQUb8CGU1in0qZmnLg9SU1WSltWRDS4HzbVlvHgu/WmdF8I547t31FNitVBVZqV7UAJ+tMveUTxuR0oXXMXSWBMaEL7aN57W64jVK/CAX1LQKZ3pmXkudA6ze5MrbWUQDrbW0dkzRlcae9uBYJCjZ7rZur4KV2VoMN7jcuAdkJRORCAYpL3LR4snvekcgHXuUgwDrknAzzkJ/6pXSt0DfBEoBjqAR7XWw4vOaQbOAm3hQ71a67cmes1Uc1XYGJuYZXpmHmtx4VV7PHdliLn5ALtaUp/OibhrWy3ffvYyL53t4T1vaknLNS52DjMwOsUv3rtx4Vi9087z170EgkFMUtOH3qEJJqbnUlYwbTnWIjO1VXau9o6l/VpidZLp4X8d+JDWegdwHvi9GOfsBb6ptd4d/i9ngj3cmKlTqL38k5HVtY2VabtGuaOY1o3VvHSuJ22lFo6c7sZutbBns3vhmMflYGY2wFCBp+wiLocXXKV7wDaiqbZUevg5KJmAv1VrfV4pVQQ0ALH2DNwHtCqlTiqlnlVK7UjieikXefwvxDx+aHXtYMpW1y7nYGsdw2PT6DRsK+mfmuVV3c/+7bUUWW48pXnCFT+9MlMHCM2/t1st1DkzU8mysaaUgdEpJqZmM3I9EZ+EUzpa69lwAP8xMAv8YYzTpoDHgL8D3gZ8Tym1VWsd1zw9pzO51YBud9myr5utoXnn0/PBFc9Nh2xcM+LS1WF8/hnuvn3dku1IVfvur7Tzj09rXmsb5PC+5pR8JoTa9/LRdubmA7zr3k03tddqtwLgm5rL2n3O5vd3sat942xZX01tzY2UTjrb13pbDf/2s3bGZgI0NyZ2nVy6f7HkevtiWTHgK6UeAb6w6PBFrfX9WuszQK1S6pPAt4GD0Sdprf8k6ssnlVJ/BmwFTsXTuMHBcQKBxNIAbncZ/f3L5xCDwSBFFhMdXSMrnptq8bQvnZ47cRXDgPVuR8x2pLp9d2x2c/SUl/cc3piS3bEi7XvqxSs01ZRSbjXf0t4yexFvdA5l5T5n+/sbbXJ6js5uH7tanAttSnf7yq2h7/GZS33UlltX/f5cun+x5Gr7TCZj2Y7yis/yWuvHtdbrov8D3qGU+vmo0x4Ddi5+r1LqN5RS0SOCBqGngZxgGAbO8sKcmnn68gC3NaR2de1yDrXWMT0zz+uX+lP2mVd7x+jsHeOeXZ6Yr3ucDrwyNZP2bl94wVX6B2wjKkuLKbMXydTMHJNo8nYW+Gul1J7w1+8FjsY4717gowBKqXsBM3AxwWumhauy8BZfDfmmuNo3nrJiafG4rbESZ7k1pXPyj5zuxmI2cde22piv17scdA9MFPyKz7au0fCCq8wM2EKoM9VYIwO3uSahgK+1ngfeB/xvpdRJ4D3AxwCUUp9SSn02fOpvAQ8opc4CnwPer7XOqYpWhTgX/9TlyOrazAV8k2Gwf3sd5zqGGB1PvpLizOw8x871cMdm15JPKR6nnYnpOUYzVNohV7V7fdS7HNhtmd3RtKmmjK5+vxSxyyHJDNoeBfbEOP7lqD93AQ8keo1McFfY8E/NMTk9l/YViLni5OVBaipLqM/QjI2Ig611PPFSJ8fO9/LWO5uS+qzjZ3vwT81xz87Y6RwI9fAhVGKhsnT1eeS1IBgM0tY1yh1RU1YzpbGmlLn5AD2DEzS4U1+OWaxeQa+0BXAuzMUvjLROZHXtzk3OjG8yXu90sKG+jJdSUEHzmZc7cZZb2brMRh6RqZmFXEStZ2gC/9RcxubfR2uslRILuabgA36kLv7ASGGkdc6HV9emujpmvA5sr+Nq3zjXkwgCA6OTnHqjf2Hj7KVUlhZTYrUUdBG1dq8PICMrbBerq7ZjMRuSx88hEvArC6uHH1pda2ZzGlfXLufObbWYTUZSg7cvnAm99+6d9cueZxgGHqe9oIuotXWNUmK1LKS3MsliNtHgKuWalFjIGQUf8MtKiiguMuVUwH/5Qm9atoiLrK5t3eBM++rapZTbi9mx0cmxcz0JrbEIBIMcPd3Nrk3uhaez5dS7HAXdw7/c5WOjpzxr9YQaa0q52jde8DOlckXBB3zDMHDn0EydkfFpvvx/z/FX/+c0c/Opnd3Q2TPGqH8mLbXvV+NAax0j4zNcuLr6UgsXOocZ9E3xwF3xDfp6nA58E7OMT+bM8o+MmZyeo2tgPCvpnIjG2lLGJmYLfqZUrij4gA+hgdtc6eFHcq4d3WM88VJnSj/75BsDGAbsbMlO/j5i9yYnJVYLL55ZfVrn6OluHDYL+1uXT+dEeFyhmUiF2Mu/0u0jGMxcwbRYmsK18SWPnxsk4MNCDz8XHjvbvKOYTQZ7t9Tw/ReuLPwCSIVTbQNsyuDq2qUUWczs21LDa5f6mZ6Zj/t9C4XSttVRHGd5hvqFmTqZC/hfe+ICT77YkbHrLeVy+GdnYzZ7+JHNUPIsj58LsSAdJOAT6uFPTs8zMT2X7abQ3uWjqbaMX36borKsmK98/9yqguJShnxTXO3N7Ora5RxsrWN6dp7XVlFq4di5XubmAysO1kZzVtgotpgythnK5PQcL5zp5rGnLjIzm/z3LRltXaPUO+04bNn7BW+3FeEst+VVD/8r3z/P33z3bLabkRYS8AF3ZKZOlksszAcCdPT4aPGUY7cV8dGHt9E7PMm//vRy0p99qm0QyOzq2uVsWleBq8LGi2e7437PkdNemmpLaa6Lv0qhyTCoy+BMnY5w3ZqxiRleysDWjksJBoO0ezOzw9VK8q02/vnOIV691E9Hd+qernOFBHyi5uJneeC2q9/PzGyAjeEiV1ubq3hwXyPPvdbFmfbBpD771OUB3JU2PBleXbsUk2FwYHsd5zuH45qR1NkzxtXe8WVX1i7F48pcEbVIkGhwO/jRK9ezlhroG55kfHI2owXTltJYU0rP0ATTWX7iiYd/apbR8dAA85MpHkPLBRLwubHatj/LPfy2hUUyN3pl7753Iw0uB1978kLCM02mZ+Y5f2WYXS3p27s2EQdb6wgG4fj53hXPPbpCobTl1DsdDPmmmcxAyq7d66O22s4j923GO+Dn3JWhtF8zlkzvcLWcxpoygsFQhybXdYdTf811Zbx2qX/NreGQgA84bBZKrGYGszxTp71rlHJ70cLWixAa4Pz4O7cxPjHLP/7wYkI9xvOd4b1rb8uNdE5EbbWdFk/5immd2bl5jp1fvlDaciIlFnqG0pvHj6RRNtaXc/j2BsodxfzoxPW0XnMp7V4ftmLzwt89myIlFq715f7AbeRJ8EMPKoosJp48trZ6+RLwidTFz/5c/Davj42eilt64U21ZfzC4Y28ovs5dm7l3vBipy4PYCs2o7K0unY5B1rruN7vX3YWx2uXBkKF0paoe7+STE3NHPJNM+qfYaOnnCKLmbfc3sCZ9sGs9BLbukZDC65M2X+ic1XYKLGa86KmjnfAT7HFxPr6Mu7Z5eHYud6sdwRTSQJ+mLsyu3Px/VOz9AxNLJlzfdudTdy2roLHfqRX9QMYCAY51TZI64b0712biDu3hkotLDfAeeS0F2e5ja3NSxdKW467sgSzyUh7Hr+9++ZpkG+6vQGL2cSPXslsL39qZo5r/eM5MWALofGade78GLj1DvipdzowGQZvC1d0ffrlq1luVerkXgTIksjiq2wNsnVE5kzXxw74JpPBR9+xjUAQ/v6J8wTibGdnzxij4zM5MztnsdKSIna2ODl2rjdm3fSBkUkuXBnm7p3LF0pbjsVsoq7avpCfTZcOrw+L2bQw97zcUcz+7bW8eKY7oyt9O7rHsr7garGmmjKu9Y3H/XObLd5B/8ITobPCxv7ttTx/yotvYm2sFJaAH+auKGF6dp6xLC3Bb/P6MID1SwR8gJrKEj5w321cvDrCj05ci+tzT10ewAB2tmS3nMJyDrbWMeqf4cKVW0stHD0Tyu8f2lGX1DXqnfb09/C9ozTXlt70JPXg3kZm5gL87GRXWq+9uB2Q3QVXizXWljI9M5/TVWknp+cY8k3jiSo099D+ZmbnAvz4lfj+veU6CfhhkYHSbOXr2ryjNLgdK27CcvfOem6/zcW//ayN6/0rPyKfujxIS0MFZfbiVDU15Xa2uHDYLLdU0AwEg7xwpptt66viKpS2HI/LQf/IJLNz6ZkaOB8IcKV3jA2LfmGvqylla3MVz77WlfLaSEtp6/JRV23P+orqaDdW3OZuWieyb0L0QHe908Edm9385NWujMzySjcJ+GE3pmZmvgcSCAbpCA/YrsQwDB592xbsVgtf+f55ZueWDiLDY9N09o5lvVjaSoosJvZtreU13X/TP6oLV4YZ9E1zdwJz7xerdzoIBqFnKD3f34U1FDF61Q/sa2R4bJpXdF9arh2tb3iCsx1DbElwvCNdGlwODCO3a+pEBvU9i0pJP3SgmcnpOX76euae0tJFAn5YpAeZjR5+b2RXojgfwcsdxfzy27dyrW+c7x1tX/K8yN612drsZDUObq9jZi5wU6mFI6e9OGwW7ticfPs9UdsdpkP7MnVrdrY4qa0q4UcnrqV1jCgYDPLYM5ewmA3eeXB92q6TiOIiM/VOR24H/EE/FrNpYY+MiA315WxfX8XTJ65lvVxGsiTgh9ltFhw2C/1ZCPgLwWIVg2y7b3NxeJeHHx67yqVrIzHPOXl5AFeF7ZYeSy5qaSinprKEF8PbH45PzvLapQH2b6ujyBJfobTl1FWXYBjpK6LW3u2jtKQId+WtqSeTYXD/3kY6uscWFtelwyu6n7MdQ/zC4Y1UleXeHr6NNaU5PRffO+CnrtqO2XRrWHzowHp8/hleOBN/KZBcJAE/iitLdfHbvT5KrOZVbyr+S/dtwlVp46s/OH9LfnF6NrR37a5NubW6dimGYXCgtY6LncMM+aY4fj5UKO2eXfEXSltOkcWMu7IkbT38UEqufMl7fWhHHXarhWfiHGxfrcnpOb7540s015bxljsa0nKNZDXVlDLom87ZvQm6o2boLLalqZKNnnKeOn415myyfCEBP4qrwpaVlE6bd5SN9avflchWbOHj79jOoG+Kf/nxGze9duHKMLNz2du7NhEHttcSBI6d7+XIqVChtKba+AulrcTjdKRlQ/PJ6Tm8A/5bBmyj2YotHN7t4VXdl5ZOxXePtOMbn+HDb1Mxe6i5IDJwm8x+xukyPTvPwMjUkk/DhmHw8IFmBkanePlC+sdi0iU3fzKyxFWZ+bn40zPzXO/zsyHBRTKb1lXw8IFmjp7p5lV9I/998vIA1mIzqqkyRS1Nv5oqO5saKvjh8atc7UusUNpy6l12eoYmUt5DuxKukLnSNMj77liHgcGzr6Z28K+zZ4yfvHqdN93RsOwvnWxrDP/yzsUVtz2DEwRh2VIUuza5aHA5ePKlzoS258wFEvCjuCpKmJ0L4MvgdmxXenwEgsGktqF716ENNNeW8Q8/vMjo+DTBYJBTbQM5u7p2OQdb6xifnMViNrF/++oLpS3H43QwHwjSN5zaHnZkhe1KwdZZYWOPcvOzU16mZlIzxS8QCPKPT1+kzF7Muw9vTMlnpkuFo5hyR3FO5vEjazSWG+8yGQYP7W+ma8DPifPZK32djPyKBmkWmYufyYHb5WZ3xMtiNvGxd25jenaerz91kc7e0OrafErnROzbWoPFbGKvcqd8444bM3VSm9Zp9/qorSqJa977g/saw5ukpCZg/OxkFx3dY/zSWzZhz+JGJ/FqqinlWg7OxfcO+DGbDGqqll/vcee2GlwVNh5/9o283BVLAn6USMDP5MBtm9dHTVVJ0gujGlwO3vOmFk63DfK1Jy5gADtyeHXtUhy2Ij79oT184IHNKf/suurQgFwqZ+oEg0Hau31x/8Juaahgo6ecH71yLekyA6Pj03znZ+1sba5KqGx0NjTWlOId9GdsEVq8vAN+aqvtKz4Rm00m3n5XE7pzGH11JDONSyEJ+FEii68ytfNVMBikzTuaVDon2n171rFtfRXX+/1sbCinPIdX1y6nua4sLatES6wWqsutKS2xMDw2zej4zKpy5w/ua6RveJLTbcltavPt5y4zOzfPh96q8mImFoRKLMzNB9MyeJ4M7+BE3JsD3b2znsoyK0/kYenkpAO+Uup2pVTMLYuUUsVKqX9SSl1QSr2mlNqS7PXSyVZsocxelLGqmZFgEc8K23iYDIOPPLSVCkcx+7clV3tmrfI4HSktonYjJRf/9/COzW6qyqxx10OK5fyVIY6d6+Wh/c0LTy75oLEmNHCbS3n82bkAfcMTca9XKbKY+bnDLZzrGOJKT35tg5hUwFdK2YEvAUt1JX8T8GuttwK/DXwjmetlgqvClrGUzsIOVynchq663Mbnf/0Q9+1Zl7LPXEvqnQ66B/0pq9rY3u3DYjYWphzGw2I2cd+edVzoHE5o5ensXIB/euYSNZUlPHygedXvz6a66hKKLKacqqnTOzRBMBj62YjXQwfXU2K18ESebYOYbA//88AXl3n9YeCfAbTWzwNupVRTktdMq9Diq8z08Nu6RimymFjnjj9YxCPRMsKFwOOyMzMXYChF3+N2r4+m2jKKLKv7p3R4l4fiIlNCvfynjnfSOzTBf3hwc0pWIWeS2WSiwZVbJRbimaGzmN1WxH17GnhN59c2iAkHfKXUuwC71vo7y5zmAaLXIncDOd31jCy+ykTd7navj+a6srybOpnPIr24VOTx5wMBrvT4Epr7XlpSxKHWeo6d71nVNODe4Ql+8GIn+7bU0Lox/wblAZpqQ5uh5MosF++AH8MIPX2sxv17GymymHjqWP5skLJ8LV5AKfUI8IVFhy8C5cD9K7zdBER/Vw0g7uF5pzO5nq/bvfpVmhvWVTJ//Crm4iJcMeqipMrsXIDO3jEePrQhoXZmQq62KyKR9lntoRozvqn5pP9+Hd5RZmYD7FY1MT9rpc9/74OK517v4mXdz/vfuvLwVjAY5K++e5Yii4lfe+9unEmWjM7W93frRhfPn+rGbC1a9u+QqfYNjs1Q73Tgqa9c1ftamp289cB6nnyhg1/5uVZqqlIzljI6HhoSrShNfT2kFQO+1vpx4PHoY0qpjwF/ADyvlIocOwnco7WOHo25DtQDbeGv6wBvvI0bHBxPeEWb211Gf//qB4as5lA6RLcPEEzjHrDDk3PMzgXwVJUk1M50S/T+ZUoy7Su3F3Gpcyjpv98r50IPr66y4ls+K572WY1QJc0fHG3n3p31K6aFTlzs4zXdx/vvv43AzFxS7c/m97fKHpqB9fr5niV3Ystk+zq8o9Su8t9hpH2HW+t48oUO/uWpC0lPJZ6amePJY508/fI19io3H3/n9lV/hslkLNtRTiiXoLX+qta6RWu9W2u9O3xs96JgD/Ak8GEApdTdwJTWOqeffzI1F193hnZ3yqVdiQqFx+WgOwVF1Dq8oQqZNUk8CT6wrxHfxCwvX1h+c/p8KI4Wr8gAdy7k8efmA/QOxT9DZ7FUbIMYCAR5/pSXP/i7Y/zgxU72bHbz7ntbEvqslaQ8eayU+pRS6rPhL78EWJVS54C/BD6U6uulmquiBFuxmYudI2m9ju4cpqrMSnW5beWTRUrVOx14ByeSziG3d4fy98nMgd/WXEWD28EzK9TK/+7zuV8cLV4lVgvuSltO1NTpG55kPhBctobOSm5sg7j6zeovXBniM984wTeeuoi7soRPf3gPn3jX9rTFhRVTOvHQWhtRf/5y1J+ngEdTcY1MKbKY2LulhhMX+/jgA5uxFqdnFsSlq8PSu88Sj8vB5PQcI+MzCdeNn5yew9vvZ89md1JtMQyDB/Y28o2nLqKvjsTcqaqzZ4yfvJb7xdFWozG8qXm2LbXL1WrUOx3codz85NXrvP2uphW3KQXoGZrgX5+9vLBnxad+bjv7ttSkfQFdfncV0uRQax3TM/O89kb/yicnwDcxQ/egn5YULbgSqxNZUZnMdLorPWPhCpnJfw/3b6ultKQoZq38fCqOthpNNaX0DU0wPZPdHaS8g34MoG6Ve1Es9nCc2yCOT87yzR9f4o+/epyLV4d5z5ta+B8fv4s7t9ZmZLW0BPwYbmusxFVhW9h9KdVSUTBNJK4+BdsddnSn7ntYXGTmzbc3cOryAL3DN68CzrfiaPFqrCklCFzvz24v3zvgx1lhw1qU3JP8+rrQNojPnLjG7Nytv8Tm5gM8c+Iaf/B3L/GTV69zz856/uyTB3hof3NG11JIwI/BZBgc2F7H+StDDI/FrBqRlHbvKCaTQXNdbk97XKsqHMXYrZak6rm0h4veparmz5vvaMBkMm7KA+djcbR4NdaGBm6zncf3DiQ+YLvYwwfWM+qf4WhUJdRgMMjrl/r5468e51s/eYP19eV85iN38uG3baHCkflaVxLwl3CwtY5gEI6dS30vv63LxwZPedK9CpEYwzCod9mT6uG3h3cpS5XKUit3bavl6OluJqZCWwDmY3G0eDnLbditlqzm8ecDAXqSmKGzmGqqpMVTzlPHOpkPBLjaO8Zf/MvrfOn/nMFkMvjtR3bxO+/dlfKV9ashAX8JtdV2WhrKeeFsT0pXBAYCQTq6faimWwfnROZ4wjV1EjHkm2JkfIYNKU7JPbC3kenZeY6c7s7b4mjxMoxQ/aFrvdlb6zEwMsXcfCCpGTrRDMPgofA2iH/xzdf5zNdPcL3fz394cDOf+cid7GxxZv0Xd0pm6axVh1rr+cenNVd7x1OWfvEO+pmamUc1V6fk80Ri6p0OjpzuZnxydtVpmXSNwTTXlaEaK/nxK9exWEx5WRxtNRprSjlyuptAIIjJlPlAmIoZOovt2uRinbuUNq+Pt97ZxDsONufU2Iv08JcR2n3J4IWz3SufHKdIsIg1/U5kjieJgduOcIXMpprUj8E8sK+RQd9U3hZHW43G2lKmZ+fpG8nchkPRIvWU6pOcoRPNZBj87vt38+f/8SDvzcGBdgn4y3DYiti9ycXx870p26Gn3TuKw2ZZmCkisiMyNTORImrtXh+NNauvkBmP3eGNsg9sr83b4mjxalqojZ+dPL53wE91uTWuefOrUW4vTnh9R7pJwF/BwdZ6xiZmOds+lJLPa/P6aGmoyHour9BVV9goLjKtejOUQCDIlZ6xlA7YRjOZDP7br+zjo+/YlpbPzyUelx2zyeBqlvL43oGJlOXv84UE/BW0bqymzF7EiylI60RWZ6YrWIj4mQyD+mrHqnv4XQN+pmfn07qGwmI2FcSeBkUWM3VOe1Z6+IFgkO5Bf0rz9/lAAv4KLGYTd22r5eTlAfzh6XKJ6uj2hVZnpnCHK5E4j8u+6pk67d5RQBbNpUpTTWlWAv7g6BQzcwEJ+OJWh1rrmZsPcuJCX1KfE9nSUHr4ucHjcjDkm2Zyei7u93R0+3DYLNRUpW+vhELSWFPG8Ng0YwlWmkzUwgwdSemIxZpqS2lwOZIutdDeNUq9055zI/eFKrL71WpW3LZ7Q4vmZAwmNSIrbjPdy1+YoeNae2scliMBPw6GYXCwtY7LXaP0DiW2HD8YDNLe7ZOCaTkk8jgfb1pnamaOrgEZg0mlbNXG9w74qSgtxlFgnS8J+HHav70OwyDhXn7/6BRjE7OSv88h7kobFrMR91z8K91jBIOSv0+lcnsxlaXFXO3NdMAvvBk6IAE/blVlVratr+alcz0JbXDe3hUa7JMefu4wm0zUVtvjTulEKmSulZr0uSLTtfGDBTpDByTgr8rB1joGRqd449rIqt/b5vVhLTLTUIA/ZLms3umIu4ff7vVRU1lCmT3zVQ7XsqbaUroH/czOpWZx40qGx6aZmpmXgC+Wd8dtbqzFZl5IIK3T7h1lQ31ZVmqGiKV5nHb6RyeZmV15I472bl/KC6aJUB5/PhBMakOa1YgM2HpSWFIhX0jAXwVrsZm9ys0rF/uYjiNARMzMznO1dzwluyOJ1PK4HASDoS3nljM8Ns3w2LQM2KZBZOA2U3l8b3h1tfTwxYoOtdYzNTPP66vY/vBq7zjzgSAt0jvMOZ44p2bKgqv0qa2yU2wxZSyP7x3wU2YvKsjUnAT8VdrcVImz3Lqq2TptEixyVm21HcNYuWpme7cPs8mgqTZ7m1esVSaTwbqaUq71ZaamjnfQX5AzdEAC/qqZDIMDrXWc6xhiZDy+7Q/bvT5cFTYqSnOzgl4hKwrXnV8pf9zh9dFUW7qmyxVnU2O4xEIqNxuKJRgM0j3gL9hqtRLwE3Bge2T7w964zm/3jkrvPofVOx14l0npBAJBOnrGZDpmGjXVlOKfmmPIl/o9pKP5/DP4p+YKcsAWJOAnpN7poMVTzgtnu1fskQyPTTPom5b59znM43LQOzSx5J4H3gE/0zPprZBZ6BozVBs/Hbtc5RMJ+Ak62FpHV79/xR/Qhe3wZIVtzvK47MwHgvQvsfNSe3dkS0P5pZ0uDW4HBnA1zXn8yJOcBHyxKvu21mI2GSsO3rZ7R9O2HZ5IjUgRNe8Sm6G0e33YrRZqpUJm2pRYLbirSjLSw7dbLVQ4Cm+GDkjAT1hpSWj7w2Pnepbd/rDN66OpNj3b4YnUqF9hu0OpkJkZTTWlXEvzXHzvQKikQqF+L5OOQkqp25VSMUdalFLNSqkxpdTJ8H9PJ3u9XHKwtQ7fxCznOmJvfzgfCHClxye53xxnK7bgLLfGnKkTqpA5LguuMqCxppS+kclV7U+wWt5BP54CK4kcLande5VSduBLwFLPR3uBb2qtP5nMdXLVjhYnpSVFvHi2h12bXLe83tXvZ2Y2IAO2eaDeFbumTmePVMjMlMbaUNrzev84TeuqUv75vokZxiZmC3YOPiTfw/888MVlXt8HtIZ7988qpXYkeb2cEtn+8PU3BpiIsf1hZIcrWWGb+zxOBz2DE7dUQo0M2EoNnfRrSnNt/O4Cn6EDSQR8pdS7ALvW+jvLnDYFPAbcAXwO+J5Sak2NlhxsrWNuPsCJi7duf9jeNUq5oxhnhS0LLROr4XE5mJkLMDg6ddPxdq8Pd6WN8gJchp9pVWVWHDZL2mrqFPoMHYgjpaOUegT4wqLDF4Fy4P7l3qu1/pOoL59USv0ZsBU4FU/jnM7klrG73emfGeNyldJYW8rLF/t5zwNbbnrtSu84W9dXU1MTu3eYifYlo5Dat7UllJLzzwbYFvW5nT1jbNvgTOhahXT/UqVlXSU9w6HAnOr2DftnKLGa2bzRlZJB21y8fytZMeBrrR8HHo8+ppT6GPAHwPNKqcixk8A9WuuxqPN+g1AOfzB8yABuzX0sYXBwnEAgsaXWbncZ/f2Zqc1x19ZavvPTNs5d6qWmKjQgND45S1f/OPu31cRsRybbl4hCa1+JORQALrYPst4d6gEOj00zMDoVKqG8ymsV2v1LlbqqEn76ehfzgSBDg6nt6bddG6Gu2sHAQPKfm6v3z2Qylu0oJ5TS0Vp/VWvdorXerbXeHT62OzrYh90LfBRAKXUvYCb0dLCm7N9Wi8HN2x92yGKdvFJaUkS5o/imgduFRXOSv8+YxppSZuYCePtTn9Yp9Bk6kIZ5+EqpTymlPhv+8reAB5RSZwnl8N+vtc7MtjYZVF1uY+v6Kl4827NQaqGtaxTDgPV1+ffYV6g8TvtNUzM7whUym6VCZsZEauN3hCvMpop/apbR8ZmCzt9DktMyI7TWRtSfvxz15y7ggVRcI9cdbK3jqz+4wBvXR9ncWEl7t48GVykl1pTcYpEBHpeDl86FfmkbhkG7d5TGGqmQmUkelwOzyaDD62PrutQ9HXdHNj0p4CmZICttU2bP5hqsRWZePNtNIBikw+ujRern5JV6p4PJ6XlGxmduVMiUdE5GWcwmGtwOTr7Rn9JSyQvbGhZ4D18CfopEtj88cbGP633j+KfmJPebZyLBwDvoxzsYrpApK2wz7t7dDVy+NsL5zuGUfaZ3wE+xxVTwU6Ql4KfQwdY6Jqfn+c7P2gBkhW2eidRI7x7wy4BtFt29ox5nhY3vv3AlZZ/pHfBT73RgKtAaOhES8FNINVdRXW7lbPsQJVYLdQW6yUK+KncU47BZ8A5O0NEdrpBZLd/DTCuymPjFN2/i0rUR9NXU9PJlhk6IBPwUMhkGB7bXAbCxvqzgexP5xjAM6p2OhR7+BvkeZs2DdzVTbi/iBy9eSfqzJqdDO2kVev4eJOCn3MHWUMBvaZB0Tj7yuOxc6xunq9/PBknJZY2t2MJb72ri3JVh2pKcotk9KDN0IiTgp1i908Hv/tJuHtzXmO2miATUOx1MTM8RCAYlf59lb769AYfNknQuv9C3NYwmAT8Ntq2vxm4rynYzRAKig4IE/OyyFVt4cF8jp9sG6exJvIyBd9CPxWzCVVnYM3RAAr4QN4nsfuWqkAqZueC+PY2UWC1J5fK9A37qqu2YTRLu5A4IEaW63Iat2Cy9+xxht1m4f886Xr3Uz/UE6+uEtjWUGTogAV+Im5gMg9/4xR28+96WbDdFhD2wrxFrsZknXupc9XunZ+cZHJ2S/H2YBHwhFtm6vhp3ZUm2myHCSkuKeMvtDbx8oZeeoYlVvbdncIIgMkMnQgK+ECLnvfXOJorMJp5YZS5faujcTAK+ECLnlTuKObzbw0vneukfmYz7fd4BP2aTQU2VPLGBBHwhRJ54+13NmEzw5LH4c/neAT81VSVYzBLqQAK+ECJPVJVZuWenh6OnuxnyTa38BkIbl0s65wYJ+EKIvPH2/U0APHX86ornzs4F6BuekAHbKBLwhRB5w1VRwoHWOp4/5WV0fHrZc3uHJggGZcA2mgR8IUReefhAM3PzAZ5++dqy58kMnVtJwBdC5JXaKjt3bavlude7GJuYWfI874Afw4C6apmhEyEBXwiRd95xYD0zs/M8c2LpXr53wE9NZYlsQh9FAr4QIu94XA72bKnhJ69exz81G/McmaFzKwn4Qoi89I4DzUzNzPOTV67f8trcfIDeIQn4i0nAF0LkpabaMnZvcvGjV64xOT1302t9w5PMB4IyJXMRCfhCiLz1zkPr8U/N8dzrXTcdl12uYpOAL4TIWxvqy2ndUM3TL19lemZ+4bh30I8B1DmlDn40CfhCiLz2zkPrGZuY5Wcnb/TyvQN+nBU2rEUyQyeaBHwhRF67bV0lW5oqeerlq8zOhXr53gEZsI3FkugblVKPAv8T6A0fekJr/elF5xQDfw/sBSaBD2itLyZ6TSGEiOWdB9fzF986yZHT3dy720PP0AStG6uz3ayck3DAJxTEf0dr/S/LnPObgF9rvVUpdRj4BrA/iWsKIcQttjRXsamhgiePdaIaK5mbD8gMnRiSSensAx5VSp1RSj2mlKqKcc7DwD8DaK2fB9xKqaYkrimEELcwDIN3HlrPkG+a7/y0DZAZOrEk08PvBj4HvAj8KfBXwAcXneMJnxf9nnXAyrVNAaezNInmgdtdltT7003alxxpX3LWWvve7Crl+y91cqptEIAdqga7rSgdTQNy//7FsmLAV0o9Anxh0eGLWuv7o875c6AtxttNQDDqawMIxNu4wcFxAoHgyifG4HaX0d8/ltB7M0HalxxpX3LWavvevq+RL10bobrcin9sCv9YfBulrFau3j+TyVi2o7xiwNdaPw48Hn1MKVWhlPpPWuvILwIDmLvlzXAdqOfGL4M6wBtHu4UQYtV23eZifV0ZrgpbtpuSkxLN4Y8Dv6+Uuiv89a8D341x3pPAhwGUUncDU1rruNI5QgixWibD4L988A4+/s5t2W5KTkoo4Gut54H3An+rlLoA7AF+H0Ap9Sml1GfDp34JsCqlzgF/CXwo+SYLIcTSrEVmKYm8hIQHbbXWR4A7Yhz/ctSfp4BHE72GEEKI1JGVtkIIUSAk4AshRIGQgC+EEAVCAr4QQhQICfhCCFEgkimtkE5mCK0aS0ay7083aV9ypH3JkfYlJxfbF9WmmPNSjWAwsdIFaXY3cCTbjRBCiDx1D3B08cFcDfhWQtU4u4H5Fc4VQggRYiZUzuYEML34xVwN+EIIIVJMBm2FEKJASMAXQogCIQFfCCEKhAR8IYQoEBLwhRCiQEjAF0KIAiEBXwghCkSullaIi1LqA8AfAUXAF7XWf73o9d3AV4Fy4HngU1rrWHvvpqt9/43QzmAAT2itfz/G6x8BhsOHvrL475BuSqnngBpgNnzok1rr41Gv7yZL91Ap9TFC22dGbAD+SWv961HnZPweKqXKgReBd2itryil7gf+F1ACfFtr/Ucx3tMEPEboXmvgg1rr8Qy17xPAbwJB4BVC3+OZRe95FPifQG/40BNa609nqH1fJ7S63h8+5TNa6+8uek9W7h+wDfjTqJcbgONa63csek/G7l8y8jbgK6UagP9BaHvFaeBFpdRzWuvzUac9BnxMa31MKfX3wMeBv81Q++4HHgRuJ/QP7YdKqV9Y9IO8F/glrfVLmWhTjDYawGageZkgnrV7qLX+KqFfNiiltgPfA/5k0WkZvYfhfZy/Qui+oZQqAb4G3AtcA55QSr1da/3Uorf+DfA3WutvKaX+GPhj4L9koH2bgd8j9O9kDPgG8GvAFxa9dS/wO1rrf0l1m5ZrX9S1D2utu5d5a1bun9b6SUJ7c6OUqgNeAP5TjLdm5P4lK59TOvcDz2qth7TWfuA7wHsiLyqlmoESrfWx8KFvAI9ksH3dwH/WWs9orWeBC0DTonP2An+olDqtlPorpZQtg+0DUOH/P6OUOqWU+vWbXsz+PYz2t8Afaq0HFh3P9D38OKGA6Q1/fSfwhta6I/xL8zEW3SOlVBFwmNDPKKT3Pi5u3zTwq1prn9Y6CJzh1p9DCJUyeVQpdUYp9ZhSqioT7VNK2cPt+Vr4e/gZpdRNcSnL9y/aXwBf1lq/EeO1TN2/pORzwPcQCqoR3cC6VbyeVlrrc5FAqZS6jVBq58nI60qpUuB1Qr2vO4BKQr2WTKoCfgL8AnAf8Cml1ANRr2f1HkaEn5ZKtNaPLzqe8Xuotf5YeD/niHjukQvwRT1Fpe0+Lm6f1rpTa/0jAKWUm1CK7P/GeGs38N+BnYSeVP4qE+0D6oBnCaXl9hMq+vXRRW/L2v2LCP8bfhPwl0u8NSP3L1l5m9Ih9MsquhCQAQRW8XpGhFMRTwC/F90zCOcfH4o67/OEUgMZy/uF0yALqZBwyuYh4EfhQzlxD4FPEsqR3yQX7iHx3aPF5xDjnLQKp0CfAv5ea/3Txa9rrX8h6tw/B9oy0S6tdTuhDkfk2l8CPkworRKR9fsHfIJQSumWgmSQvfu3Wvncw79OqCpcRB03P4at9HraKaUOEepB/39a639Y9FqTUuojUYcMbgycZqp9dyul7lumDblwD4sJ5cf/PcZrWb+HxHeP+oAKpVSkRnl9jHPSRim1hdAg5D9orf97jNcrlFLReWkDyNTA/A6l1LsXXXvx9zCr9y/s54FvxXohm/dvtfI54P8YuE8p5Q7nAd8N/DDyota6E5gKB12ADxHq4WSEUqqR0CDjB7TWsX5QJoE/V0ptCA+e/hrw3RjnpVMl8BdKKZtSqgx4NLoN2b6HYTuBS+FxmsVy4R4eB5RSalM4IH2ARfcoPIZzBHhf+NCHF5+TLuHv6zPAH2mtP7/EaePA74cHLCGU9snUfTSALyqlqsK5+k8svnY27x+AUspFKKXYscQp2bx/q5K3AV9r3UXo0f054CTwTa31y0qpJ5VSe8OnfRD4glLqIlDK0vm3dPhdwAb8L6XUyfB/n4q0T2vdTyhV8X1C08wMYKl/kGmhtf4BoXTT68CrwNe01i/l0D0E2EioF70gx+7hFPDLwL8B54GLhAcXlVJfVUq9K3zqrwKfUEqdJ5SnvmXqZpp8DKgF/nPUz+Fno9untZ4nNMb0t0qpC4Rm9Pz+0h+ZOlrr08CfEZr9ch44GZnpkiP3D2L8DEa3L5v3b7WkHr4QQhSIvO3hCyGEWB0J+EIIUSAk4AshRIGQgC+EEAVCAr4QQhQICfhCCFEgJOALIUSBkIAvhBAF4v8BEH98onkVdSwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(res.fx[0:res.total_num_search])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "688ca138",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res.export_sequence_best_fx()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "id": "157d202b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([-2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667,\n",
       "        -2.30707667, -2.30707667, -2.30707667, -2.30707667, -2.30707667]),\n",
       " array([9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394,\n",
       "        9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394, 9394]))"
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res.export_sequence_best_fx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "142d7d20",
   "metadata": {},
   "outputs": [],
   "source": [
    "res = physbo.search.discrete.results.history()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "id": "0db59dfc",
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "'history' object is not callable",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "Input \u001b[0;32mIn [173]\u001b[0m, in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[43mres\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n",
      "\u001b[0;31mTypeError\u001b[0m: 'history' object is not callable"
     ]
    }
   ],
   "source": [
    "res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cc1ee569",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}