{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "unable to import 'smart_open.gcs', disabling that module\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import numpy as np\n",
    "import sys\n",
    "\n",
    "from sklearn.linear_model import SGDClassifier, LinearRegression, Lasso, Ridge\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn.decomposition import PCA\n",
    "import seaborn as sn\n",
    "import random\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from collections import defaultdict\n",
    "from sklearn.manifold import TSNE\n",
    "import tqdm\n",
    "import copy\n",
    "from sklearn.svm import LinearSVC \n",
    "\n",
    "from sklearn.cross_decomposition import PLSRegression\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "import torch\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.svm import LinearSVC\n",
    "\n",
    "import sklearn\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "import random\n",
    "import pickle\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import cluster\n",
    "from sklearn import neural_network\n",
    "from gensim.models.keyedvectors import Word2VecKeyedVectors\n",
    "from gensim.models import KeyedVectors\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "def set_seeds(seed):\n",
    "    \n",
    "    np.random.seed(seed)\n",
    "    random.seed(seed)\n",
    "    torch.manual_seed(seed)\n",
    "    \n",
    "    \n",
    "def load_glove():\n",
    "    with open(\"glove_data/data.pickle\", \"rb\") as f:\n",
    "        data_dict = pickle.load(f)\n",
    "        X,y,words_train = data_dict[\"train\"]\n",
    "        X_dev, y_dev, words_dev = data_dict[\"dev\"]\n",
    "        X_test, y_test, words_test = data_dict[\"test\"]\n",
    "        X,y = X[y > -1], y[y>-1]\n",
    "        X_dev,y_dev = X_dev[y_dev > -1], y_dev[y_dev>-1]\n",
    "        X_test, y_test = X_test[y_test > -1], y_test[y_test > -1]\n",
    "        return (X,y, words_train), (X_dev, y_dev, words_dev), (X_test, y_test, words_test)\n",
    "    \n",
    "    \n",
    "def perform_purity_test(vecs, k, labels_true):\n",
    "        np.random.seed(0)\n",
    "        clustering = sklearn.cluster.KMeans(n_clusters = k)\n",
    "        clustering.fit(vecs)\n",
    "        labels_pred = clustering.labels_\n",
    "        score = sklearn.metrics.homogeneity_score(labels_true, labels_pred)\n",
    "        return score\n",
    "\n",
    "def compute_v_measure(vecs, labels_true, k=2):\n",
    "    \n",
    "        np.random.seed(0)\n",
    "        clustering = sklearn.cluster.KMeans(n_clusters = k)\n",
    "        clustering.fit(vecs)\n",
    "        labels_pred = clustering.labels_\n",
    "        return sklearn.metrics.v_measure_score(labels_true, labels_pred)\n",
    "    \n",
    "    \n",
    "def load_word_vectors(fname):\n",
    "    \n",
    "    model = KeyedVectors.load_word2vec_format(fname, binary=False, limit = 150000)\n",
    "    vecs = model.vectors\n",
    "    words = list(model.vocab.keys())\n",
    "    return model, vecs, words\n",
    "\n",
    "def save_in_word2vec_format(vecs: np.ndarray, words: np.ndarray, fname: str):\n",
    "\n",
    "\n",
    "    with open(fname, \"w\", encoding = \"utf-8\") as f:\n",
    "\n",
    "        f.write(str(len(vecs)) + \" \" + \"300\" + \"\\n\")\n",
    "        for i, (v,w) in tqdm.tqdm_notebook(enumerate(zip(vecs, words))):\n",
    "\n",
    "            vec_as_str = \" \".join([\"{:.6f}\".format(x) for x in v])\n",
    "            f.write(w + \" \" + vec_as_str + \"\\n\")\n",
    "            \n",
    "            \n",
    "def get_inlp_p(ws, rank):\n",
    "    ws_normed = ws/np.linalg.norm(ws, axis = 1, keepdims = True)\n",
    "    ws_normed = ws_normed[:rank]\n",
    "    P = np.eye(ws.shape[1]) - ws_normed.T@ws_normed\n",
    "    return P\n",
    "\n",
    "def get_relaxed_p(rank2adv, rank):\n",
    "    \n",
    "    adv = rank2adv[rank]\n",
    "    U,D = get_svd(adv)\n",
    "    U = U.T\n",
    "    W = U[-rank:] \n",
    "    P = np.eye(X.shape[1]) - W.T@W \n",
    "    return P\n",
    "\n",
    "def get_v_measure(X,Y,P,num_clusts=2):\n",
    "    X_proj, Y = X@P, Y\n",
    "    \n",
    "    #before = compute_v_measure(X[:], Y[:], k = num_clusts)\n",
    "    after = compute_v_measure(X_proj[:], Y[:], k=num_clusts) \n",
    "          \n",
    "    return after\n",
    "\n",
    "\n",
    "def bolukbasi_debiasing(directions, input_dim, rank):\n",
    "    \"\"\"\n",
    "    the goal of this function is to perform INLP on a set of user-provided directiosn (instead of learning those directions).\n",
    "    :param directions: list of vectors, as numpy arrays.\n",
    "    :param input_dim: dimensionality of the vectors.\n",
    "    \"\"\"\n",
    "#     import scipy\n",
    "#     if np.allclose(directions, 0):\n",
    "#         w_basis = np.zeros_like(directions.T)\n",
    "#     else:\n",
    "#         w_basis = scipy.linalg.orth(directions.T) # orthogonal basis\n",
    "\n",
    "#     svd = sklearn.decomposition.TruncatedSVD(n_components=rank)\n",
    "#     if rank < len(directions):\n",
    "#         w_basis = svd.inverse_transform(svd.fit_transform(w_basis.T))\n",
    "#         w_basis = w_basis / np.linalg.norm(w_basis, axis=1, keepdims=True)\n",
    "#         w_basis = w_basis.T\n",
    "    pca = PCA(n_components=rank).fit(directions)\n",
    "    vecs = pca.components_\n",
    "    #P = np.eye(input_dim) - w_basis@w_basis.T\n",
    "    P = np.eye(input_dim) - vecs.T@vecs\n",
    "    return P\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "(X,y, words_train), (X_dev, y_dev, words_dev), (X_test, y_test, words_test) = load_glove()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_all, vecs_all, words_all =  load_word_vectors(fname = \"glove_data/glove.42B.300d.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "projs_rlace = {}\n",
    "projs_inlp = {}\n",
    "\n",
    "for seed in range(5):\n",
    "    \n",
    "    with open(\"interim/rlace/run={}/Ps_rlace.pickle\".format(seed), \"rb\") as f:\n",
    "        projs_rlace[seed] = pickle.load(f)[0]\n",
    "    \n",
    "    with open(\"interim/inlp/run={}/Ps_inlp.pickle\".format(seed), \"rb\") as f:\n",
    "        projs_inlp[seed] = pickle.load(f)[0]\n",
    "        \n",
    "specific_words = [[\"woman\", \"man\"], [\"girl\", \"boy\"], [\"she\", \"he\"], [\"mother\", \"father\"], [\"daughter\", \"son\"], [\"gal\", \"guy\"], [\"female\", \"male\"], [\"her\", \"his\"], [\"herself\", \"himself\"], [\"mary\", \"john\"]]\n",
    "vecs_specific_words = np.array([ [model_all[p[1]] - model_all[p[0]]] for p in specific_words])\n",
    "vecs_specific_words = vecs_specific_words.squeeze(1)\n",
    "projs_words = []\n",
    "for i in range(1,11):\n",
    "    P = bolukbasi_debiasing(vecs_specific_words.copy(), 300, rank=i)\n",
    "    projs_words.append(P)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{1: 0.5066666666666667,\n",
       " 2: 0.5077551020408163,\n",
       " 4: 0.50421768707483,\n",
       " 8: 0.506938775510204,\n",
       " 12: 0.506938775510204,\n",
       " 16: 0.5048979591836734,\n",
       " 20: 0.5046258503401361}"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open(\"interim/rlace/run=1/Ps_rlace.pickle\".format(seed), \"rb\") as f:\n",
    "    data = pickle.load(f)\n",
    "    \n",
    "data[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_pca(X,y, path, title, method = \"pca\"):\n",
    "    import matplotlib.pyplot as plt\n",
    "    plt.rcParams['font.family'] = 'Serif'\n",
    "\n",
    "    if method == \"pca\":\n",
    "        pca = PCA(n_components=2)\n",
    "        M=6000\n",
    "    elif method == \"tsne\":\n",
    "        pca = TSNE(n_components=2, learning_rate=\"auto\", init=\"pca\")\n",
    "        M=2000\n",
    "        \n",
    "    X_proj = pca.fit_transform(X[:M])\n",
    "    ax = plt.axes()\n",
    "    ax.set_axisbelow(True)\n",
    "    ax.yaxis.grid(color='gray', linestyle='dashed')\n",
    "    ax.xaxis.grid(color='gray', linestyle='dashed')\n",
    "    \n",
    "    y_text = [\"Female-biased\" if yy==1 else \"Male-biased\" for yy in y]\n",
    "    plt1 = sn.scatterplot(X_proj[:, 0], X_proj[:, 1], hue = y_text[:M])\n",
    "    plt.legend(fontsize=19)\n",
    "    ax.set_title('{}'.format(title), fontsize=25)\n",
    "    ax.figure.savefig(\"{}/{}.pdf\".format(path, title), dpi = 400)\n",
    "    plt.show()\n",
    "    plt.clf()\n",
    "    \n",
    "def get_maj(Y):\n",
    "    \n",
    "    from collections import Counter\n",
    "    c = Counter(Y)\n",
    "    p,q = list(c.values())\n",
    "    return max(p/(p+q), 1 - p/(p+q))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAEPCAYAAABfmE8WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOydd3gc1bm43zM721crrbplWa6YZkIxxoAB0wKhBBMMMe0S30C43EAIJBATwi8NSLkEklACARNKcDChg4FQQgvFNhhs3AD3omKturbvzJzfH7M71korWbJktIJ9n2ee1c6cOfPNjPabM9/5ipBSkidPnjx5Ri7KcAuQJ0+ePHkGR16R58mTJ88IJ6/I8+TJk2eEk1fkefLkyTPCySvyPHny5Bnh5BV5njx58oxw8oo8z4hGCDFTCNEkhHhKCCFGWv+7gxBinBBCdluOHW658gwfeUWeZ6RzNlACfAsoHoH97w61wAGpJU8e1OEWIE+eQXIPcDDwHyll8wjsf8BIKZPAKoAceUnIM8yIfGRnnjwjFyFE+gd8nJTyzeGUJc/wkTet5MmTJ88IJ6/I8+xRhBDfEkIsEkLsEEIkUp+LhBBndmuXdQJPCLG/EGKBEKJOCKF12dbvCT8hhE8IcaMQ4jMhREwI0SyEeEMI8e0+jttn/33sd2BqYjSYOtYaIcSPe5soFUKUCiG+K4R4RgjxqRAiLIQICSFWCSFuEUJUDuX9yPPlJK/I8+wRhBBOIcQTwFNAKfB94IjUZxnwtBDiMSGEI7VLtgm8GcAzwJvAGcB5QGMf7bPJUQ4sBm4A6oFzgJOA+4EbgZu7ND851d8H/eg/2/ajUv0+Cpza5Vz/APyml37+kNpnXEqeY1Myvgn8EPhECLFfX+eYJw9SyvySX4Z8Af4KSOAjwNVtmxtYntp+V5Z9ZWppBcZ323a9+W+btf2xWfp6IbVtMWDvtq0UU7mn9x/Xy7n02n+37fVAabdtp6S2RYCCLPs+CGwDCrNsuzS174d9XOc+ZcsvX40lPyLPM+QIIfYFvpf6eqOUMtZ1u5QyCtyU+nqZEGKfXrpaIKXc1G3d7cD4fsoxDXNkDHCzNL09usrRBNzRn776ySOpPrvyOmBgPrymZtnnGeAKKWV7lm0PAAlgan5Unqcv8oo8z57gHCBtE/53L21eTX0qmL7a2Xin+wopZUhKubmfcszq8vcbvbT5Tz/76g8fdF8hpYwDaeXew94tpXxGSvlsts5SD5761Nd9h0rIPF8+8oo8z57ga6nPJillR7YGqRFoS+rrgb30ExykHOlRbJOUMtRLm4ZBHqMrvfmZR1OfrmwbhRDfFEI8IYTYnJrs1NILMDbVzDeEcub5kpEPCMqzJyhMfUb7bGXajYu7tO+OPkg5/P2QI9nHtoEyYHmFEA8Ac4E24FbgPXZO6AK8DFSx8w0nT54e5BV5nj1B2t7r2UW79PZs9uGhIP020Jcc9j107F0ihJiFqcQB5kgpX8nSZigfNHm+pORNK3n2BJ+kPkuEEFlH26n16dwlK/aQHGu6yFHQS5vh9NM+IfUZYuecQZ48AyavyPPsCR7HdIkDOLGXNielPo1U+z1B10nE43ppc/QeOnZ/SP/+egsWsmOaVfLk6ZO8Is8z5Egp1wL3pr7+PyFExiRf6vvPUl/vllJ+tofk+AB4MfX1+pRi7CpHOlBpuEh7zHgxsyt251KG0fSTZ+SQt5Hn2VP8EDPgZjbwlhDi98BmzAjGeZieKo8DP0rvIISY0q2P8UKIJiAhpfy864aUUt67l/bbpZRtqXX/jel6OB14RQhxK1CH6c53A2Z2wxu7C9+f/vuQd5OUMiyEmAw42KmMR6f2Scv3OPBdzLeTh1Pb3ki1Px34HyAOOLvuC4T7ee55vioMd0RSfvnyLpgmg7MwR8WNmB4ijZjRlt/K0l72smzO0nZcH+3ndmtbgBmAtA5TMTZjeoOcAEzosl/VQPrvY/uxqe2bdyUfptL+EWYEbCQl32bMiM/9s/QxdyDnnl++Gks+jW2erzRCiEOAZZiug15pBvDkyTOiyNvI83ypEUIcL4S4tI8mh6Y+P8wr8Twjlbwiz/Nl5xjgDiHEuO4bhBAeTFs+mME4efKMSPKTnXm+CjiAN4UQ/4eZdTEOTMG0Te8H/J+Uck+5QObJs8fJ28jzfKlJ5SP/FqYXyGSgAjPSMwi8j+n+2Ftirzx5RgTDoshLS0vluHHjdtkuEong8ewqynv4ycs5tIwEOUeCjJCXc6gZbjmXLVvWJKUs67FhOFxlpk6dKvvDX//61361G27ycg4tI0HOkSCjlHk5h5rhlpNeiowMy4j80EMPlR9++OEu29XW1jJ69OgvQKLBkZdzaBkJco4EGSEv51Az3HIKIZZJKQ/tvj6nvVaCwcGmo/5iyMs5tIwEOUeCjJCXc6jJVTmHTJELIdxCiE+EEH8Yqj6XLl06VF3tUfJyDi0jQc6RICPk5RxqclXOoRyR3wR8PIT95cmTJ0+efjAkilwI8V/Au0D3QrmDYurUbLVqc4+8nEPLSJBzJMgIeTmHmlyVc9ABQanq3vtKKa8XQnytj3aXYqblpLKyknvvNbOcHnbYYZSVlfHCCy8AUFNTw0knncT8+fPRNI2VK1cyd+5cnnnmGRobzQpYs2fPZv369axYYdYjOPLIIykoKODll18GYMKECRxzzDE8+OCDAHg8Hi688EKeeOIJWlrMMpFz5sxh1apVrF69GoCZM2eiqir//rfpUjx58mSmTZvGggULAPD7/Zx77rksXLiQjg6z8MwFF1zABx98wJo1a1i2bBknnHACmqbx1ltvAbD//vszZcoUHnvsMQCKi4s5++yzeeSRR4hEIgDMnTuXt99+m40bNwJw8skn09nZyXvvvQfAgQceyKRJk3jyyScBKC8v58wzz+TBBx8kkUgAcMkll/DKK6+wdetWAE477TSCwaD1Gjh16lSqq6tZsmQJy5Yto6qqitNPP5358+djGAaKonDJJZewaNEi6urqAJg1axbbt29n2bJlu7xPAA6HY8ju09q1a63jDuV9+vxzM4HiUNynadOmsWrVql3ep0X/ehWnv4RktJNvHHd0v+7Ts8+aadSH4j6VlpZy8MEH75H7NJS/J03T2GeffYb8Pg3170nTNIAhv0/9/T31SjZXloEsmHmlfw5cB7wGvA1c1dc+effD4SEv59DRHxk/re+Q/3X/Yjl23iJ53C2vy3fXB6VhGF+AdDsZCddSyryc/YVe3A8HPSKXUt6c/jtVMMAnpfzTYPvNk2ck0x5Nct2Tn/DxNjM1+MamCP/9wAcs+sFR7FXRW9W5PHl2jyHLtSKEmI2ZoMghhDhPSvnoYPusqhoZVa7yckJ7eztNTU3W6+lgOPTQQ1m7du0QSDX0GFIipWla6UvGpG7w/YPdcLA7Y33nji2sbbHtaTEtcvladuWrKKfNZqOgoIDi4mKcTueg+srpgKA8I4NYLMbWrVuprq7G7XYjRNYSlCMaKSXhuE5DR4ykbhDwOCj22nGo2ZVyXNNZvyOE3u33NaHMh8+5++MnTTeIawa6IXHaFZy9HD9PbiOlJJlM0tHRQWtrKzU1Nf1S5iMyICht+M91vupyBoNBysrK8Hg8Q6LEczHoIpbU2dQcJpLQSOoGjZ0xmkIJehsIOVUbo4oyR+MBjwOXuvs/uaRuUNsWZUMwxObmMOsbQ0QSWp/75OK1zMZXTU4hBA6Hg9LSUgKBgDVpvLvktCI3DGO4RegXX3U5Y7EYPp9vj/SdK8SSRg+l3RJOkNR7v6ZFbjsTy3yMCXgYX+plVKEL1bb7P7loQqc9mrS+64akvj2GbuQzmI5k/H4/nZ2dg+ojp/ORK0pOP2csvupyapqGqub0v9KgyXbpbIro8w1EUQRep4p3cOZPi2wPjVhCRzckNuXLZ876qmC329F1fVB95G3keQbN2rVr2XfffYdbjD1KQtPZ1BQhru38wdUUeyjyOL4wGUKxJBubwhnritwOSn0OErqBU1Vw2m0oX8I5ii87/f0NjUgb+aJFi4ZbhH6Rl3NoaWtrG24ReuBQbYwr9TCm2MOoQhdjAy78LvuQHyehmeaT9kiCWDJzlOZ22BhV6LIUtcehUuBWWR8MsbUlwvrGEJ2xZMY+uXgts5GXc3Dk9PtwOioq18nLObQkk8ldNxoGnKrN8hIJBoMoXteQ9h9L6mxuCpNImVBsimB8qRePQ019Vyj1OfG77JYb5IZgyNpfArWtMdx2FUdqUjVXr2V38nIOjpwekefJk6vouk5lZSXFxcUIIXC73VRWVmZdHA6HFd7eF50xzVLiYE5mNnfzjBFC4LTbcDtUkrpBd8OoZhgYu2kuffjhhy1598ScR/f+Y0mdzliSWFJHsfXuRvmd73yHmpqaEeHZ8uSTT+7Ra9gbOa3IZ82aNdwi9Iu8nENLUVHRcIuwS0pKSmhoaOCpp54CzFwjDQ0NWZcjjzwSoGtai6x0tb+niSV1etvFoSoIMu3hHoeKatu5biDX8qKLLqKhoYHDDjus3/sMhO79r28MsakpzLodIXB4e30ANTc3097ePiTBZoNlV9dz9uzZe/Qa9kZOK/Lt27cPtwj9Ii/n0JILP9hdMRAZJabr4KamMFtbIoTjWoZCjyd1QvEkhW67ZRJJE/A6UHrxSHHZbYwt8aCmXGo8DpXRRW7r+0Dl/KJIK+z0p0RS1xYjnszuyvn888/T2NiYExWEcvF6Qo4r8nSmsFznqyRnUjdYVdvOok/qeH9DE22Rof/HTmeyy2UGIuPTz7/EkaecRSiu0R5NsjEYJpIwR9+dsSTrGkNsDIbZ3Byh1OfEbbchhKDM56SwjwlVIQR+t529KnxMrihgfKkHtyPTRJGT1zLLwFsi0XqJcxBCDDqEfajIyetJjivyPLnH62t3cMad73DFPz7mvPuWcNMLazHyASlZmTt3Lr/4xS9oiWqWv7lhGDxy/90cNvUQiouLqa4s4ztnncK/X1qElJL6thiji9xc/7//xdf2qsFht3HJJZfw7LPPcuihh+L3+5kyZQpPP/00AG+++SYzjz6KitJiDj7oIJ5//vkecrz//vtcfPHFTJ48mcrKSgKBAMcffzyvvvrqgM7HMAz+/Oc/c+CBB1JcXEwgEOCYY46xZOkvae/IhrrtXHPZXE6ePoUj9hnDN085ia5uyZFIhMrKSgoLCxFC8Mgjj2T08+ijjzJr1izGjRtHWVkZFRUVnHvuuaxbt67HMZubm7nyyiuZMGECo0aNYvz48Zxxxhk89NBDPdquXbuWc889l/LycoqLixk3bhxXXHFFVht9bW0t559/PoFAgKqqKo455hgWL148oOsxFOS0Iv+i7Uy7y1dFzvr2KNc/vYquevuJZdv7jG7cHbxeb8b3Zz6uZcbvXmf8dS8w43ev88zHtUN6vN2hu4x90dWOfd0Vl3DH/93EtdffwNbaBl778FMOPeIofnTpf/HI/LuRSHRD8vxzz/H+++8DsHjxYhYvXsy7775LQ0MDY8aM4ZxzzmHhwoX8+9//5o033qC+vp5Ro0Zx1llnsWXLlgw577rrLhYvXswLL7xAQ0MDW7ZsYfLkyZx88sm89NJL/T6P888/n5/97Gf86le/oqmpifr6eo499ljOOuss/vSn/ic8FUIgpeQXP76C7/zPD3h16Wo+WP4JsWiUmTNn8vHHZqExj8dDQ0MDt956a9Z+rr32WpxOJx9//DHBYJDFixezbds2jjjiCGprM/9HZs+ezX/+8x/eeecd6uvrWbFiBYFAgIsvvjij3dKlSznssMOIRqOsXr2alpYWnn32WV5++WUOP/xw4vG41batrY2jjz6aN998k9dff526ujoWLFjAL37xC7Zt29bv6zEkZMttu6eX/uYj3759+0BS9Q4bXxU5P2vokGPnLeqxLFu+cogkNInH49bfT3+0Xe5zw0sZx9vnhpfk0x+Z5xJJaLKhPSo3N4VkWzguk5o+pLLsSsY33nhDAtLlcsmKioqMxeVyyV/84heyPZKQK7a1yt/dOV8C8rKrfiLD8aSMJzW5qrZNrtjWKvc94EDpcrnlfz7ZJCNxTUop5bp16yQgKysrZTKZtI79r3/9SwKyuro6Y/1LL70kAfmHP/whQ87rr79evvDCCxnyJ5NJWVZWJmfMmNHj3GbMmCFtNlvGukcffVQC8uc//3mP9occcoh0u92yubm5X9duxowZEpAPPPiQDMWSMpbQZDwelytXrpSAPPbYYzPa33fffRKQf//73zPWn3TSSXLHjh0Z65YvXy4B+bOf/cxa19zcLAF57bXXZrRtbGyU+++/v/XdMAy57777Sr/fL1tbWzPaPvfccxKQP/7xj6118+bNk4D829/+ltH2008/lUKIHtewL9asWdOvdvSSjzynR+Tp6hm5zldFzlF+F1NrAhnrbIrI8JIYCtrb262/b3n5M6LdAmOiSZ1bXv6MeFJnUzDMjo4Y7dEkW1oitEa+GD/frjJCdq+VOXPmAOBzqkwo9fLqc0+Ybb/1Tdx2Gw7VRk2xB0UIDjhoKrFYlHXL38dpz/xZTps2LcOVbcyYMQAcfPDBWdenK9uk5bz55ps59dRTM/pUVZVJkybxySef9Ot802aNU045pce26dOnE41GrWpA/eUbJ5+E16nitNtob29nypQpjB49mrfeeovW1tZd7v/yyy9TXl6esS4dHdn1vLxeLwUFBdx///0sWLCAWCwGQFlZGatWrbLaLVu2jLVr13LEEUf08E6ZPn06kBlUlzYpnXTSSRlt9957b+tefFHkdEBQntyiwG3nN2dN4WdPr+LDLa2UFTj5zbcOQKV5jx2zri3a6/poUu8xQdbYEcvq/dEdKSXRpE4saWAT4HLY9lhKWEUR+Fx2tmxcD8C3zz6rR96bRCKB1+ulo6WxR4h9UXEJmm5YCbccDjMtQGlpaUa79IRgOJwZxh8MBrn99tt54YUXqK+vtzxmWlpa+p1I7bPPPgPgjDPO6FX2+vp6AK644gqeeOKJjDYff/wxo0aNsr4rikJFRUWP44wePZra2lo2bty4y/qYq1ev5rbbbmPx4sU0NTVl5L3pOinpdDp57LHH+O53v8uFF17IZZddxte//nXOO+88zjzzTOx2e8Y5vv3221RWVvY4ntfrzchSmC4n1/W80lRVVfUw7+xJclqR19TUDLcI/eKrJOfelX4e+O9pNLTHKHCrVPrdrF07uBSc3UkrKoCqIje1WZR5VbcUsWn6O+0ajmtsaoogU3s4UyH4/VXmXWXsjd6CgN5++2322WefXe4fiidTnzobgiGqAx68XXKZ9ycJWjKZ5JhjjqGxsZF//vOfnHDCCZbiOuqoowY8Mdcf2e+8807uvPPOAfWbvp7p0fKu0iEvX76cI444gsmTJ7Nw4UKmTJmCEAJN06zz68opp5zC1q1befHFF3nqqad46qmnePrppznwwAN5/fXXKS4uttqeccYZLFy4MOtxu7+JgXmNh9urJqdNK91fWXKVr5qcBS47e1UUUOnPrkwHi9/vt/6+9uS9cdszlavbbuPak/fGZbf1yPpXXuDEvgtTj24YNHTELSUOZjBO2iVwoDL2l7QCzDYRpmkar732Gs3N5ttNPKlT1xrrIp/B1pYISW1gE8vvv/8+GzZs4Dvf+Q7f+MY3siq5riQ03UqL2zU97kBk7w+GYVhFhWHn9ayvr0cIwfjx4/vc/29/+xuxWIyf//znHHDAAbtU/LquY7fbmTVrFg899BD19fVcdNFFrFixgrvuumuX5wjw+eefW6N2MItSQ/bUF+m3ky+KnFbkX/WCDUPNSJGzqamJeFKnLZLg6L1K+dU396OqyIUARhe5+e1ZB3DmwaNx2W1MKPVS4nPidaqMCXgIeB27/FEbMntKWE3vvxtlU1PTQE+LCy+8EIDHH3+8x7Znn302w2yR0HuG2id1IyOEvz/0NsKVUmbY0gEiCY31jWFiqQjTHR0xtNTxBiJ7f3nttdesv5uamli8eDHBYJCZM2cSCAT62HPnCL77eW3evLlH2/Xr11NcXJwRhOXz+bjqqqsALHv8IYccwn777ceSJUt6XBvDMDjnnHMyXCDPOussoOfc07p163rsv6fJaUWe56uJ0+21oiBr26LsW+Xn5auOYdPvTuPd647nzIN3Rvi5U9GME0q9BLwO7P0o3KAqgmJvT9OIx7Fny6bNmTOHOXPm8OCDD/Lwww9bOajffPNNvv/973PjjTdaCkzNEs2pCDHgvOOHH344Y8aM4aGHHuKtt94CTOV+3XXXZYw8dcOgri2WMefQFIpbE80Dkb0/CCH405/+ZPmN79ixg6uvvhqPx8Ntt922y/2//e1vo6oqN954o2Wr3rFjB9///veztu/o6GDevHnW/EFnZyd33nknqqpy3nnnWTI98MADeDwe5s6da12ftrY2LrvsMsLhMJdddpnV57x58xg/fjy//e1vWb58OWBGT1966aUZppovhGyuLHt66a/74QMPPNCvdsPNV13O/rpO9YekpstgR0yu2NaasXze0CE1fehcC+NJXda1ReTK7W1ybV27bIvEpa4bGXJ0RhOyI5qQ8aTWY/+GhgZZUVEhA4FAhvvh4Ycf3usxDcOQkXhC3vanP8sDDzpIBgIBWV1dLWfMmCH/+c9/ZrT9znfmypKS0lTfbllSVi4/3bRN/uAHP5ClpeZ6t9stKyoqZGNjo7zyyit7rF+3bp0MBoPys88+k7Nnz7bk3WeffeQf//hHedhhh0lAVlRUyN/+/vfypj/eLUvKyqVqt0tAlpSVyx/+6MeyNRyXGxo75bbmkLztT3+WB+1C9t546KGHZEVFhbTb7dJms8k1a9bIU089VY4ePVp6PB553HHHyaVLl1rtw+GwrKiokH6/XwKysLBQjh071tr+yiuvyKOOOkoWFhbKiooKOX36dPnMM89IQDocDllRUSHfffddGYvF5D333CNPOukkWVNTIysrK2V1dbU888wz5eLFi3vI+emnn8rzzz9fVlRUyMrKSjlx4kR5+eWXy9raWhkMBjPabt++XZ577rmysLBQVlZWysMOO0y+9NJLlotlRUWFvPLKK3d5bQbrfpgvLJFn0AxlYYm2SIJoQicYimest9sUJpX7+jXiNgyJZkhURfSapwRSBXB1iSLIKMGW0HS2tUQJp+phqjaF8SXeHuHvA8GQkvZIktq2KIaUOFSFmmKPlaI2G7ohiSV1NN3Ariq4VFvG+UQTGp0xDc2Q+F0qHofa5/n2hW5ItjSHCcV31gC12xSKvQ52dOy01dsUwcQyHy57vujzUPKlLizxzDPPDLcI/SIv59BgGJKmUBxblsnKkn6aTaIJjS0tET7f0cnWlgjRPiYwhRA4VKVHHc1wXLeUOJiV65tC8QybdX/8nLsST+psa41YfSQ0g20tUcsGbRiSSEKjPZIgHNes8m1ep0qhx9FDSUeTOhuDYRo6YjSF4mxsCtMZ71mIub9y2hTBqEKX5bYpML83dWY+UNMPl6FmoNdzuMhVOXPa/bDrrHYuk5dzaBACVEWhNZxkVKGb1nACzZAEvPasJdVkyhc8HNcAgcdho7YtaimajliSmKYzscwcyaffPnc1GRrLkk42ktAxpLR8vDWt7+r13Ulk8TaJazpJQ2JTJK2RRIabZWWhi1Kvs9cRdiSuoXd7m27siOFz2rB1mXQciJxuh8rEMh8JzcCmCBQBCPrv0zkIBno9h4tclTOnFXmerxZCCMoKnGxsMqM1/S4V1aYQ8DiyBvhEEzobm8LWKFcRggq/i/r2nQoxoRnENYNIQqc5FEdVBCU+Jx6HrVeFbpo7MkeifreKbTdrYSY0HQlU+F1ICc2hOLqUqDYFVRHENYO69ljGPjvaYxQ4Vdy9mF6y5SlLVw0aDHabkvHmU16QeT1tiujhDppn+Bm0IhdCKMDzwBLAAUwEviulzB6SNwBmz5492C6+EPJyDh0eh40JpR6iCQMhwONUe7XHtkQSGeYOQ0qiCR2nasso0mCk7L9p2qMaE8u8eJzZ//29DhtlPidNoQQSic+pYlcUkrrEoZrKvL8eGvGkzubmMHFtZ/m2cr+LHR0xxgTc2G0KCS0zPzmYg2Ctj6ySXqcNgcjwhS8rcPYwEw3EkyQbAY8du03QFknisisUuu0494AiH6ycXxS5KudQ2cjfl1L+Wkp5A+ABzhqKTtevXz8U3exx8nL2jiElej/DwCGVGS8Zp7TASUkqN3dvZPP7Ns0fO7+X+Zy0RzPzr0hkVnvyThlMJVpW4KTc70IIQV17lGhy5z7xpE5c03dZVq0zrllKHEwbczSps1e5D1/qQWK3KRnFIABsKft9b7jtNiaUefG77LgdNsYEPFmLQaf9yHcX1aZQ5HEwrtRLZaG71zeEwTJYOb8oclXOQd8VKaUB3AQghFCBauCz7u2EEJcClwJUVlZy7733AmZq1bKyMsupvqamhpNOOon58+fT1NTE2rVrmTt3Ls8884xl4509ezbr169nxYoVABx55JEUFBTw8ssvA2bE1THHHGOFSHs8Hi688EKeeOIJK1fCnDlzWLVqFatXrwZg5syZqKpqJf6ZPHky06ZNY8GCBYAZeXbuueeycOFCOjo6ALjgggv44IMPeO+991ixYgUnnHACmqZZ/rr7778/U6ZM4bHHHgOguLiYs88+m0ceecTKBTF37lzefvttyxf25JNPprOzk/feew+AAw88kEmTJvHkk08CUF5ezplnnsmDDz5oVSu55JJLeOWVV6wghNNOO41gMMjSpUsBmDp1KtXV1fz73/9mxYoVVFVVcfrppzN//nwMw0BRFC655BIWLVpkRanNmjWL7du3W8Uo+rpPBx10EE1NTZSWltLa2oqmmfm33QWFBDsTxDWDQrcNv9OGTWBdP6fTic/nsyICFUWhpKSEzs5OolHzhS4QCBCLxazvBQUFgOkH7Hf76Oj2uypyq9hsCrGEhqqADYPGSLYHiaSpqQkpJW63G5fLZU1kudxeOmLJjMhGgERSpzXeTlJxEAwlMAwIeFQK7NLM19KlD1VVTdkTPR8Ymm6Y906q6FqSSCTC6EIfDalrZbcpjPI7iIVDOETvZToAACAASURBVAsLM/Jgl5WV0dbWZhUBHl1UhKbphMMdtEYkXq8XVVWtUHJd1/F6vVYAkxAi4z5lu8Y+nw9FUXZ5n1paWix/8r7uE5jXxuPxWL8/m81GcXGx1YemabjdbiKRiKUsu/fR/T6l+2hubrZyxpSUlBAKhax0s36/H8MwCIVCWftI36f0/wKYOWw6Ojqs31dhYSGaphEOh63/bYfDQVtbGwB2u52ioqI+71NRURGJRML63Xe/T9FoFMMwrKA9h8ORVe/1xpC5HwohTgauBpZIKX/RV9v+uh/ee++9XHrppUMi357kyyhnQ0eUdTtCSCnZq6KAUYW9h+Nnc52KJXXWN4YyRqxlPieVha5dTjYGg0HKysp2KaNuSDpjSRpTnhXlBU4KXPYeQTMd0SSbu5hWFCGYUGZWp09qZgSloggSqb+dqkJTKEFTNxfICWU+pJRsaspMSlXhd1Hhd2Wsk9I0eoRiWsaxi1KmieZQAkVARaELf0pmTTfQUt4q/fHQ6Q/9vZbDzVddzsG6Hw7Ze5KU8mXgZSHEw0KI70sp/zLYPtNFa3OdL5ucG4Ih/ufhZawPmqOYcSUe5n/nUCaVF/T7WLFkT7NDUzhBic+BI0tiqoRmkNB0M1NgQf+OY1MERR4HBS419T1T+UlpuspJJGOLPXTENFRFUOix41JttEUS1LWZ2RJjWtr7xTR11BR7SOoG7dEkNkVQ6XfhttsIdsZxO2wUuuwY0jTDJDQjIzthJK7RHE6Q0AzKCsyHV2NHHCHMOpsNXV4jtrVEGF/qpcBlNyc/h9j87PP5hrbDPURezsExFJOd+wHjpZTphAObgAmD7Rd2vlrlOl82OV9e1WApcYDNzRGeX1HH1V/fu1/7x5N6Vq8Kc6DcczQeTWhsbo5Y+U+KPQ4q7Ea/R6XdFXiaSMqrJf3WWeBSqfC7cag2InGNHR1xSgscCATN4Z2j76Ru0BpOUB1wU+l3IQTWw8dlVwA1QxkXd3GN7OpJU+i2E02aSaiqA25URVDf3tPG2hHVKOijNudgGGj+k+EiL+fgGAqp4sDFQogbhBC/AvYF/jAE/Vo271znyybnh1t6Bj0s2dR3qtqkbhBL6kQSGhuCYWJJHUc3RVzpd/XITGgYkoaOeEYSq5ZUdGf3dglNtwJodkVSMwjFNTNqU1EQCDpjGqG42W9CNwh47TSHElkDXMIJjUhCpzWSJJY0rOM6UmaXrrREElYyq2jqTcSp2lBtCjtSATtbWyI0hxNZH052NfOaRBM6rZEEoViScFyz+twd0nbuXCcv5+AYisnODQyRl0qe3ODk/St4/dPM4KHTv1bVa/t40syXLSX4XSqaYdAcSlBW4EBiKmGfUyWaNNjUFKbE57QCV2KaTiSLB0nXLH+xpM6OjhgdUQ2nXaGqyI3XYcOQkkhCJxTXcNgUfE4VmyJojyYJdsYpdNup9LuIJXVUm0I8aVgPDLtNob49hqbLrN4hXqdKbVvUCuQpK3BadvBsSlVPedCkVbLfrdLcTeG3R5OMLfYQimtWH6pieqy0RRL4nCpxzbxGpT4nzXGNSEJDACU+J2UFziGznef5cpHTAUHpfL+5zpdNzpmTyznvsBoe+2ArEjj7kGqO27s8a9tNTWGawwnKNAO3w0YipdAkksbOOIoQ+F0qTeGEZYMOxTVqij24HTZ2dMTwOFU6Y5kugs6UcjWz8kWtHCCxpM7mpjCTyn2E41pGNKTDpjA6YBai8LvsxDQjI2dLkdtuufylJxcl5oMo4HHQFkkgMZV4gUvF51TRDTMfS0LTSaY8SpyqkuFSqAiBzWYWFHY7zBzpApE1OCehG5T7nShCoBvmhGhdKv9KdcBDezRpvbWkHzoSMxOh12GjMEuEa18Md8GD/pKXc3DktCI/5phjhluEfjHS5TQMyebmMC3hBJWFLqoDHn55xn58d8Y4hIAxAU+vQSBbm8OWPTyeNPAX2DOUspFSbh0dmR4gwc44FX4XnTGNUYWuVASmjgBKu/iPJ3VJOK5T4XelRqMyZe7QM+zUYCrJtJkk/ZDoSls0SWmB+UN0qAqFbgdt0QRt0SQeh0q534XHYSOS0Elo0hzVe+wIzMIOTaE4xV4nNcUeatuiRBI6TlWhxOtkS3OE8gInAY+DCaVeIgmdIo+d1sjOUbmqCHxOO6oi6IwnqesmXziuYROCUp+TUFyjwKXisCkEO81I0EhCp9CT9Tb0Sq5OznUnL+fgyOn3tN5KZeUaI1nOhGbwzPJaTr39P5x9z/t88453eG9DE40dcdbUd/DvtY18tLXVGk13J9Alr7eR8hIp9ZkjTiWllIQQPcwRitiZmbChwxxtlvtdlPtd+F0q7dEkLeEEUkrGlnhoiyTZ3hphe2sUp2qGkfcR+Nh7ehC58/gVflPxCgSaYeC220jqBoYhaeyM4VQVK2FWNKnTHE6wuTlsujd2Bjlx6j5M33ccZX4XU6oDvPPhCmKajtuhUuIzTTGjCt247TaKvQ7Gl3q5ft41OOw2RpcWcfwhe/Pog/daosWSOh6nmS8mff7BUNx6+LjsNq644goqKysRQnDsscf2fgFSDKRqDzDg/gHi8Tj7778/J5544oCO1ZWByjkY5s2bN+BzTPNFyjkQclqR59nzbAiGuObxFcSS5mt8ayTJW58FufTvH/LDhcv57Uufct59S3hxZfbSVZPKfJb7H5h2YK/DxsQyL6MKXXTGkhnJptKUFThxqQpFbgdSSloiCRpTI9Ta9hi1bVG2t0bojGm0RhIZIfct4QRJ3aC0W3EIRQjcDtU0W+iyx2SrWbl+5zqn3cbogJu9K31MKvPhd9tpiyQRKbOH16n2iApN6mbelvJRVby27FNuvfdhwKwgc98dt2aUYnOoCmUFTiaW+Rhd5KazrcUKhPvW7HN4a/nnnDd3p19/ic+RNduglKbHjdepcuedd9LQ0JD1XgwFu9O/ruu0tbXtVtWk4eD3v//9Hr2Gw0FOm1Y8ngG+Rw4TuSRnXNMJxTSKPI4egTFer5ftrRHW1HWQ1A0mVxQQ7Ij1GNl6HDbW1ndmrLv5xbUctVdpj8AgT8qePLbMh6YbOGwKkaROfXMERQgCXgehmEaF32k9LAIeu5WWdVSRiyKPnaRu4FAVQnEtw4vEnsVLxDxPg1KfE5tN0BpO4lTN3NkOm2BimTcVTarSGjE9Pwpcdkp8jh65SBQhMvzaCz12bKnkW2ra1t1tfB/XDARYE4+l5WY1+H89+yRbN/+KA/bLLE6cfvO49dZb2Xvvvfn444+xKeYxInGdhG7gc6q47basbxKqTTDG5+khe3/4ItzlPB4PmzdvHtSxctWtrzu5KmduSpUiXScw1+kqp2FImjrjRLKEZu9pVte1c/Vjyzn9jnf41XOr2djFFxzgyJNm8V/3L+XSvy/j8n98zJl3vYtqU+hPUr/OmJY1FSuYytDnVCnyOIhqOnVtUdNerek0tMco8jho7IzjUBWqilz4XHZLudltCn63PZWRUKUzlnndNF1mLejgUBUUxbTLF7rtFLrtRBI6saSBbkjCKT/xApfKmGIPRZ6eEZ/dMQyJAGrbouzoiBFO6BT7Mkf9TtWGppv288pCFwpmOPVF37scXdf54x/+j3BcY1tLhC3NYfONxJC0tLRw//33c8011wDmZKvfZafYa2dUoYtirwO3Q6WsIHMyTRECr8O2W0oczJD1LwK73Y7NtvvRTF+UnIMlV+XMaUX+xBNPDLcI/SIt59bmML/716ec8uf/8N0HPuDDzX37Xg8l21sjzP3bB7y4soH69hgPL97CT59aSUcX08Crn2zNCC8PJ3QeWbKFO849yFJyLrvC1LEBy2skzTlTqxlVmBmG3h095XbYnUhCozrgpsTr6DV4B0zlVuJ14HWo1sRhcyhOideRYSYpdNtxqQqxpEGBy3Q5TOoGTrvpTbIl5bMdTmhsbYnQGUvSFIpT2xrNWnQZTJNJOKGxvTVq5VhpiyRQFcGYgIcit51Sn5NCt0pL2Jx8bA4lGB1woyoK1/3oB5SUlPDII4/wzsdraI0kaI8m2dQUJpzQuO222zj77LOpqtrpxulQFXwuO16nytNPPs6ZZ57JtAP24fiD9+L4gyfzsx9cgtayfUCJqtra2rjmmmsYP348xcXFlJWVcdZZZ/Hxxx/3u4+uvPvuuxx99NFUVVURCAS44IILMirEv/HGG1RWVuJ2uxFCZBQ/jsVi3HrrrcycOZMxY8YQCAQYP348V199dQ9/7JaWFlavXs3ZZ59NTU0No0aNYq+99mLu3Lm88cYbPeR65ZVXOP744ykuLiYQCHDAAQfwu9/9zspt0pUlS5Zw7LHHUlBQQE1NDXPmzKG2tna3rkc6V0yukdOKPFcvWndaWlqIazp/em0d9769kWAozuJNLVx4/xI+a+jcdQdDwMZguEd5tCWbWtjaErG+b2/rGVX4WUMnMyaV8sIPjuKRiw/jxSuPZsakUh65ZDpHTCimvMDJ/x47kcuPm5Q1tL4rAtFr0eDtrdFeR/Rd8aTc9zpiGg5VoTrgxvPZU0x+9AgOmD+OKf+cQdmm59gQDJtVcWIa8ZQHSzShI6XskeyqKZQg4HFQ5LETSWipCjyZfuqbmsKE4z0Dg0z3SBuFKQ+Uxs64Zf5w2c1JVyGgrLiIH/7whySTSebf9ceMPjbWNnLPPfcwb968Xs973rx52Gw2li1bRlMwyAdLl9DWtIPjZx6dUSS5L9ra2pgxYwbPPvssTz/9NC0tLbz99ttomsaRRx7JO++8069+0mzevJkbb7yRhQsXUldXx6uvvsprr73G0UcfbSni4447joaGBubMmdNj/4aGBq655hpOPvlkNm7cSGtrK4888ggLFy7k61//upXoCswcJscffzwej4c1a9ZQX1/Pq6++yurVq/nVr36V0e9f//pXvvGNb3DUUUdRV1dHU1MTN910EzfddBOzZ8/OSAn84Ycfctxxx6HrulXd/uqrr2bu3LkDuhZp0knCco2cVuQjifq2GM8sz3zKx5IG6xq/GEXuyhLUYlNExsj6a5U9E1+dPXUMAa+TfUb5OWqvMiaU+RBCMG1cMffPncYLVx7NT07emzHFu54HUBSzMIToEoavKmbld92QxHahyPVUlKfTbtanDHbG6fjgUWyLfojSsR2BROnYjuulq/GvexopzaRZHqeNQredlnAia0UdgZkTJdgZZ0tzhI3BMI2dcTTdTJLV1BknltSzmpicqg2bMIspdE2p67AplPpMM4jEfBh8//LLKSjw8+w//8GO+p3/Cw/edzennXYa48eP7/Xc99tvP+644w6r+vr48eO55557aG1t5a677urzuqW54YYbWLNmDX/5y1846KCDADOT3wMPPIAQgh/+8If96ifNtm3buOuuuxg9ejQAhx56KL/85S/ZsGEDt9xyyy73dzgcnHLKKVx//fXY7WYKghkzZnDDDTewdOlSXnnlFavt0qVLaWxs5Oyzz7Zc/MaNG8dNN91ETU2N1a62tpYrr7ySI444gl//+te4XC5sNhuzZs3i8ssv5/nnn2fRokVW+6uuuopoNMr9999PZWUlAIcffjjnn3/+gK5FrpPTijzbUz4XmTNnTuo1uecrsGcQBXsHwl6VBRy3T2ZWtstmTmBcidf6ftQ+Vfx61v74XSp2m2DukeOYdVDvEZuelM02na2wsSPG9tZIr2HyCc3MKzK2xGPm8k7lFA+mPDGyjda7ktTNUPikZtAWNYtGlC75HULLrFGi6FEqP/w/Ct12Al4Hwc4ESV1S4XfhyJLbu7LQRXMoQTQ1iSoxbdyRhI5h7MxNHk8aFHQpNiGEYHSRC9Wm4FBt1BR7mFjmY0Kpl/GlHgRmOTlNN/h8RydtuoP/uex/SSYSPHD37QCEQ538/f6/8tOf/rTPc3/xxRczzC6AlQ3vk08+6XNfML1m/vGPf+ByuTjuuOOs9YFAgJKSEiZNmsRHH32UYRbZFePHj2fixIkZ60455RQAnn766V3uX1VVxYsvvthjfbbzSh/nhhtu4I033rBG6yeffDIPP/yw1e7xxx8nkUhYcnRl+vTpAJYiDwaDvPvuu4wbN47JkydntN1dV8lcLSyR014rq1atYsaMGcMtxi5Jy3n9Kfty3VMrrfUHjC5k31H+L0SGgMfBb848gI+2trK+McQBows5qKYIe5cR+YZPV3HRjBmcuG8FSd2gqsjdr5DvSELj5VUN3PziWjqiGhceXsPFR41ndGDnKD0dcZnQDRQhqCx00dAes/zHvY7eK/2kUYTA77ZbaWkB7KG6rG3toTocqmI9JMBMvlUd8FDqc5DUJZpu4HGq2G1KRjFlS2ZNx+cyvW46Yxrt0SR+t2oFH3kcNpyqgiEl8aSBxExxq9oUwpEoUSPToyYc15jz3//DX+66g6cefZgfXTOPV558lOOOPY4x4ydR3x6lpZv5K83atWu59dZbWbx4McFgMCPVbzqHdV8Eg0FaW1tRFIXq6mprvZQSIQTRaBSv10ttbS2jRo1i2rRpGSabMWPG8MEHH2T0OWrUqB7HSY/ON2zYsEuZwLRl33nnnaxdu5aOjg6EEFae767n9bWvfY3f//73/PKXv+T444+nsrKS008/nYsuuoijjz7aavfZZ2apg1tuuYU777wz41jp3OvpvOBpGbs/IHtb1x9isVhOBgXltCJfvXr1iFDkaTm/eWAVNcUeVta2M6rQxcFjA33m8R5qRhW5Oa2o9+Ol5azqo002Vmxr4+p/rrC+/+3dzRS47Fz99Z2jnI5Y0sqPYkhJUyhOVZGLpC6x2xQKnGrGQyUbijBNRHZFWKH+SV8VjlDPiSm9YHSWyj9mdKdNCHwuBd1Q6IxpeBxOXHZbj6Amh82cHPU4bCBNV8dYUiehGRR7HdhtZkBQsDNOU8i0jbvtNsYUe0gaMmuFIkdBgIsvuYQ7br+d+++4haeeeopnF73IxmAYiaQ1JXPXt5qVK1cyffp0Jk6cyIIFC9hn3/0RioJDVXaZu707gUAgw0e6t/zZ3ZV2f0kXbOiPXPPnz+d73/se3/72t3nrrbcs5fnmm29mvDWAWVjhJz/5CZdeeilPPvkkTz75JA888ADz58/nvPPOY8GCBRnHvPnmm7niiiv6JXP6wTEURKPRnFTkOW1aGWl4nSpHTirlf2ZO5IyDRjMmkDv+5QMhmtD4aEsrz62oZdmWFpZlyYb4zw+30dxldBlPZppbEppBSzhJPKnjcdh2qcSBVG6UBOVdijQ0HPoTDDXzwSNVN8ljb0DNokwMKTGQ1LXF6IxplPqcRJM65QXODJNLoduOZkjqWqNsb43SEkmwoyOGalOwKYJo0pw4jSR0gqGdE5zRpE5TKI5qU3GoZmHorgrGblP48TXX4nQ6ueeeezjssMMYPXHfHr7oXZOCPfDAA0SjUX56/fWMnrgP65vCrGvsJNjZ/7Ji5eXlVrWcbCP4pqYmXnvttQFN1mULmklXkOpP3p6//MUsSXD77bfvcgQspcQwDIqKirj44ot58cUX2bBhA4ceeiiPPvoob775JgD77GP66Pc2Abx06VLWrVsH7DTXpGXuykBMTCOBnFbkM2fOHG4R+sWXSU7dkDy+bDtn3f0eVz66nPPuXZLVJDKuxJPh3+13qQh2powFKHSrVBS6dmlSSWMYkkhCo6kzTqXfRXmBC+ch55I45Y/oBdVIBFpBNZ0n3UZo8rco6hbZmZ5YDXbEKS9wYlMEG4Ih6tqiNLTHqCl2M7rIzZiAx0qI1d3k0hIyJ0y3NEeIawaaYVDiMysPpdW1y25je5s5X9AaSSAAv8uOQDC6yM3YMdWWV8QNN9zQw4sGyEio5XCY55HQJXVtpvujbkg+XPV5v64bmCPk9AReV7fddA76m2++meuuu25Avt4bN260ShCmee655wD41re+tcv90+fVffTe1UUxzbPPPstpp52WsW7s2LFcdNFFAFZptnPOOQen08mTTz6Z4fUC0NjYyDHHHGOZX8rKyizPluXLl2e0TZd0HCi5WnsgpxW5qua05cfiyyTn5qYwN7+w1vqeSJkWJpbtnDR1qgpXf31vPF38m10OG1VF7pSN2UwsVeR24OzFZTGh6bRFEgQ7Y4RS9THTijimme6EjZ0xQjGNxH5n03LpR3TOC7LtoiUEx5+BIaE5ZCb5Sqd4LfU5TSUozXJpHVHNUr7RpM6OjjhCCCJJndq2aNYoSokZFCSlaWMPx005dUMyqsiN32U3g6O6jKgNKSn02JlY5rXSFdx9991omsb06dMp8fXMWNg1VcA555yD3W7n97/9Dds2bwKguSnIzdf/aFe3K4ObbrqJ/fbbj5/+9KdWzVdN07jvvvu47777uPXWWwfUX3FxMd/73vesEe2HH37ILbfcwsSJE7n22mt3uf8FF1wAmJ4j6dqUn3zyCb/+9a+ztn/11Vf5xz/+Yb01bNu2jb///e9UVVVxwgknAKZt+/bbb2fjxo1cddVVVj3PjRs3Mnv2bE488UROPfVUq88//vGPuN1urrzySusNY+nSpT3s6yMeKeUXvkydOlX2h7/+9a/9ajfcjBQ5Fzy6UK7Y1ir/+cFWuWhFrdzUFOrR5qMtLXLsvEUZy4SfviAXb2iSL6+ql09/tF2urm2ThmFY+6xavVpuawnLFdtarWVzU0h2RBMyqek9jpHQdLl+R6esbY3IhvaobGiPyvZIQnZGEzLYGZPrd3TKVbVtcnNTSDa0R2V9W1Su2NYqm0MxWd9u/t2Q+lyxrVWuqm2TK7ebS11bRNa1RWRTZ0xuawnLhvao3NwUkiu2tcpPtrXK+raItV9ta0SuqWvPkHt9Y6fcFAzJTcGQXN/YmbFt1fY22RqOy5Xb2+RHm5tkSVm59BcWSUVRZFl5hbzovy+Rta0R2RlLSr3L9dF0Q7Z0RmVpWbksLApIQLpcLllRUSEXLFggpZTytddek9OPmCEL/H5ZUlYupxw0Vf75/gUSkHa7XVZUVMi3335bXn755bKioiJj/RNPPGEdq62tTf7kJz+REyZMkKWlpbK6ulqeccYZcsmSJf36H+na/8yZM+WiRYvktGnTZGVlpSwqKpLnn3++rKurs9q//vrrsqKiQrpcLgnI0tJSec4551jb//KXv8gpU6ZIj8cjq6ur5Te/+U15xx13SEB6vV5ZUVEhNU2TK1eulDfddJM8/PDD5ejRo2VFRYUcP368vOyyy+SWLVt6yPnqq6/KE088URYXF8uqqiq5//77y9/85jcyEon0aLt48WI5c+ZM6fP55OjRo+Vpp50mP/3004xrmL4Pu6KxsbFf7QbKmjVr+tUO+FBm0alDVnx5IOSLLw8PTy9Zx4+f+dzKrXLQmEJunDWFA6qLrDZNoRiz7nwvI8+3z6my6AdHMa7U271LAD5ZuRoCo3usT0+qpv2twYz+DMU1pISG9pg1slVtCmOLPWxuCuN22HCqNiIJDc2Q+N12mkNxHDYzQKi2LUaRx06wM56RVTE9QRlJ6BmpdAMeh5U/vKrQZaW3LXCb5pJwXCea1PA6VFSbQmNnjDHFHra19LQ11xR7aAknrPzoaSr9roy0uuk6nF1Jy9o9gVgas0xcyDLF2IRgfKpI9O7yVS9qPNR86Ysv7wm6+37mKoORU0rJ6rp2Vmxvx6kqHDSmaEBFjvtLcyjOrW9uy0iQtXxbO+9taKam2ENzOMH7G5ppaI/xyzP248WVDTz9cS3jSz38/PT9M5Rxmk/rO9gQDFEpDUb5nLSEExmKVaaKEyc1A7uqmHloQgmaQ2YR5oDXYdmK0yHt40q9bGuJEIprOFWblau8wu9CYHqNVBW52N4aNZNOJTRUm4LPaUMgSBpGjyIVbZEEFX4XihC0x5KMLfGapek0HVVRKPLYkdLM6xLXdCaW+bApZhre7ul345pBgctOXDMyaoxGu5WLC3bG8aYSg6XpTYGncTtsTCzzWUnDXHZbv+cXesPl6jutQq6Ql3Nw5LQinzZt2nCL0C8GI+eyLa2cf98Sa2Ra5LGz8NLD2adyaP3Po0k9a+HfUFxjXWOIqxZ+nBHC/7uzDmBimZfGzjjznlzB7EOqufy4SfhSo8zlW1u56YW1fLillfvOGIUnmqTC77SO4VAVEppOSySJ2256rcQ1ncaOGH632UdjRxyJtHKDh+Omh8vYEo8Zzi/MZF0tYdN9LJ1gy+dUGV/qJaEZuO0OdnTEaQnFrQCm7gpYYo7469qiVo50p93Wo1iGajNt/WA+YCv8Lurbd76ZeJ0q8aSOoggmlHpJ6hJFgUhcp649M2gp2wRnfxgK5d2VXMrM2Rd5OQdHTk92LliwYLhF6Be7K2dSM7j37Y0ZE2dtqXzgQ01ZgZPTDqjMWKcI02yi6TqHTSjhiAklVpj6H1/7nIllPsoLnHz/uEkUex1sSGVTNAzJB1taM4o0J3WDaEKnwKVS6LZT7HXQEkkipbTMGrqEMr+LQredHR0xyyXPkGZZuGKvw6pCv6UlwpbmCFKaJpqAx4FhSNNLRAhL4W1tjRJOaEjMgsm1bdEek4tuu41QTEsVRVZ2OTIG09Oi2GtnQqmXUYVuqgNuCpxmTvDyAhcdbS34Up466QnS7tc7W7qAL5qRlK9oJJCrcub0iPzLjiZlhi06TfcSZkOBU7Vx3qFVJHXJy6sbqPS7+M6R45hY5uW1T4Ms2dhCTYmH/3faftz++joE5qhydV0H9e0xjppUYqXm1aTB5i5ZFNNEkjplKRNL12AdRUAkbiarEpC1so9umKPb+vYYUpojU7fdTKDV2BFHUUilCzDb2hRBQjN6pAtI6gZeh4rmMVPZep2mv/eOjhhCCCoKXQhhRqvGkgaKMEfrUkpcdht2m2KlCrApAq9TxefqXSFHk6a/fGWh2yqi4XfZv7DUDHnyQI4rcr//iwlvHyy7K6fbbuOiI8Yy78mVGetP2Cd7oWMwldjnOzrZYwmVPgAAIABJREFU0hymwu9iUrmvx6Rab4TqNzN9fAWzDqoiljSV1XMr6nluheleVtsWZcW2Nr539ATGl3q57qmV1qTe8m1tJHXJ+FIfK2vb+Vp1IQuWZPbvd9lRhMiwFztsCg6bwpaWiGVT9jlVaoo9xJIGoXiSSEK3qsm7HTa8TrOIs1NVaOiI4XOqOO02OmIaAY+dTU1hynsZ8aYLPowOuNEN04UwrhmMCXhSo3iFzrjGlqZwlyyGNnxOlcZO0/+8sSNOQjdT5Ppdpimn+7HS/thmPU+dhvYoHqeKTVFojSQp9vZ0OdQNAynZ7dziu8NgcoR/keTlHBw5rcjPPffc4RahXwxGzhP3reCX39S5+60NuO02rjlpbw6u6T0xzxufNnL3m+s5cb8K3lnfTHWRixP3LUeX0BxOoACTK/1ZFcnXjzmczxo6eHXNDmKawf6j/Cz6JDPqLZLQrUnE7p4Zf1+8hYDHzu/+9RmnHlDJ2VOreeqj7YCZS8VpV2gOmwUXkpphVZQPJTRKvA46YprVr8ep0hSKW5XtfU4VVRHENXOysshjtq8qdNMeS9IWSeB1qFYmxc3NYcYUeyj2OGjpUuC4rMCF066gG5IdHTHLvu5UFWpKvOiGpL4tluFDHkvqFKbk2NIcsezrLeEEhmFW6OnuOZLOUuh22FAVgZYqZgFmsequytowJKG4xo6OGLohKS1wUui29yvPzWBJy5nr5OUcHINW5EKIicBNwEdANdAspczu8T9AFi5cOCKU+WDkLPE5mTtjPKd9rQpVERnFjLvT0B7lty+tZdZBo/n9vz7D67Dx45P25oZnVrOmoYPDxhVzSE2A+e9s4uYzp1BR6Ka2LcKn9Z3YbQrt4SjtMYM/vGJGDH7v6Al4HGqGwlYV0zb8+Y5Qj+PbbQoNHWZY/osrG/hadSHXn7ovpb4IFX4nG1PmlkhCp8httxJRpSkrcKIZBgnNzHCopBS33+00J0d1U4l7HSp+l4pmSBo7YmgpW0xHLElSNyjxOahri9IZ1fA4bYxxekjoBi67Dd0wqG2NUuSx41LNgs4CM3go2BEzHzJZsjdKCQh6eKm0R5MEvD3feFpaWiguLsZltzGhzGfKphn43fYeSj+S1NncvNMUVdcWRaTu/Z4mLWeu81WWcyhcwIdiSFAMLJRS3iKl/CFwrhBi6hD026OKSK4yFHKWFTj7VOJg2o8Pn1DC3xdvAeDCw8dy5xvrWbyphY6oxmtrG3lhZT0eh8qK7e1sDIa4cP5Svr/gI1bXtfODx1byzoZmvlZdCMALn9Rx4eE1Gcc4b3oNd7+5gVjK3t2V/505kRe6FGH+ZHs7d7y+nogGSU3LyEPutNt6JLVqDiUoSnmsOFWF0YUu/G67VaFeSyXYKvSY7n221Ei3K9Gkbk1WCgFNnXG2tUZwqQq1qbwpNkVQ3x6jrj1KY0eMYGf8/7N33uFxlFfb/832Lmm1q94lq8ty77iCO2DAdEMcaiC0hBIIhCSkQBIghBAgfh1iggOmm2ZjG1fccC+SrN57b9vb98dIK68lwA3Q+72+r8vXZc3OPHt2npkzz5xzn/sQZlBicXr6kpiDz7NEYFDCEkAiGZo2eLJmiUouJUyvIjpEg141uKXcqW82IDa78Hi/vdHGuWK4NkI4Ff+X7XS5XOccsjnnFbnP5ztVRk0CDM6EXcAZo7XHwbG6Lpq67MSbNCilAjEhaj9PWimX+EMH/civ72Z2ehj1XXYEAcYnhJAQGk2oRsHTV2ZT22HjopRQQjQKdpW20ml18o8bRnOstovoYDWJJi1r9lVT2NjDT2el0NrroLnHwew0MyPC9TzzeWHA983JCONEk434qF5Meh2tPWIZ/FDa494+YnmIRoFSJqG+y+5fHavlogJhWF9Cs9fpQSUfvM4QEHB7RalclVxCh1U83uERdVEE+sr8T4rTe30+Oq2uvlZzAiadAh/QYXEik4rt5dotLkw6BSq5NODYML1yUNs7p9uDRKWj3eJAKZOiVki/kQkz1LmQS4WAB98F/N9Fd3f3OWu4nNcYuSAIVwAbfD5f4RCf3QHcARAREcGKFSsAmDBhAmazmc8++wyAuLg45s6dy8qVK/F6vaxatYrly5ezdu1ampubAbjqqqsoLS3l6FFRWnXKlCno9Xo2bNgAiMps06dPZ9WqVYDI/Vy2bBnvvfeenz507bXXkpeXR35+PiAKSslkMr+YTmpqKuPHj/dTCw0GA9dddx1r1qzxr8BvvPFG9u/fj9frZcWKFcyZMwe328327dsByMrKIjs7m7fffhsQ42tLly5l9erVfoW65cuXs2PHDr840bx58+jp6aGkqo5NbcGsK2j1n8P7ZyeTFqHnilHRvLW/Zsj+l/1OQyWT8HleI8lhOnRKGTvL2mi3OLhohAmby0tauJ6loyNRdFRQd+Agj91yCxs2bMBmUzAlKZTtJa08v6mYUK0Co1bB5EQjwUofzyzJ5Lkvymi3Olk8MpJZaWZaOnvFirdwGdHBGqR9sd9T+dxquRSdQkqXx0u33R0Q4rC5ROoigljR6PH6sDnFxsonr+xNegUqmajr0ngyL97Xf50N5nD3PzgUMgk9FisquYxQtQSdVExiChKBOKMau7WXCK0CF6KeuUouQeJx0tHeQ2hoKL29vfgECU1WX4CzjwlWopL4kMoV2KxWXE4HMpmMkJAQWltbkas0/jg6iA8jo0aG2+3C7XZjsVj816lCoaCzsxMQGxoHBwf79bVBFILq7Oz096YMDg7G6XT6ryetVotMJvNrm8jlcnw+H62trX3nR8BkMtHR0YHbLb4phISEYLfbsdlEBpVOp0Mikfivc6VSiU6no62tTZxXiYTQ0FDa29v9K9RTx+h3TP1aKCqVCo1G47//pFIpRqPRP4bP58Pj8WC1WrHb7UOOoVarUalUfgGt/jHa2tr8Alr989Qvt2swGPB6vfT29g45xsnz1B/iMJlMdHd3++Vvg4KC/PPk8/mwWCznPE/9fPSmpiaampro6ekhPDyclStXAqLg2FB+7+tw3kr0BUGYBVwBPODz+b7xnfF0S/S3bdvGzJkzz4t93yX67WzoslHVZvVX6OmUZ/+cPFDZztJX9wRsU0gl/Gv5OHw+H/srO2jsEpN5mwub/fvcNCmeJJOG3eXtfSJOcv6+pQT7STKz981JQaeUMTHRSG7s4MTqhvxGnt1QRElzLwqphGWT4mi3uLhhQiwZUQbeP1hLTIiGj47UkV/fzVVjopF7HaQYfOjkPrw+sdxcp5Jhcbhx9TFQdCoZHRaxqrPT6sJ1ip63RiFSDtstTn8yUquQopRLcHt9SPtWvQ63F5/Ph6KvjN/l8WHSKfoqS0XVxS6b6KR0ShlOjwenu7/sXYxNK05DVnco2JxiGOhk9AtlWR1upBKxOYZSJg1oHefyiLkBn08slpJLhTPWGj8b2O32YVuNeDL+L9oplUrR6/UYjUaUytPLl3ynJfqCICwCLgLuByIFQYj3+Xx7vuWwb0VxcfEP6sjbehwUN/fi9nhJDtN9bUOG4uJiwlJHc+vrB/yVjTdMjOXnF6dh0n/7BHm9Phq77cgkgl+Le6i4qtPj5WBVBy98UUK4Qcn8rAiWTUrhuglxVLZaCNHKidCr+PXHeWTHBPPFiSbmZkYEOHGADw/X8ZPpyeREBwds9/l8YpcclYycmCDmZ0fgAz7Pa2ReVgS7SlsBgeyoIH727hFq2m08c2UOJxq6eedALYIAy6ckEGFQ8eTH+SikEuZlhRMXquVYbSe5sWpe2lLF8snxqBQqXt0eKJH6+MJ0lHIpT36cH7A9I1JPkknLqLgQ/rjuhL+sXxDgj0uyCQ1WMiLJxKHqdt7aV4NB5SMlLJhXt5dx46Q4XtxcGjDewpwI/nrtqK9VZvwmvHeghoc+PqlFmVnHmPhg3j1Q698mlQh8ePcURsYEDzXE94r/LTpAF+w8N5wP1spY4G3gALAV0AL/AM7Zkf+QqGm38vN3jrC/UnwFiwpW8e/l40kbonTeHBHF85uKA0rg3/yqhoszIpj9DZxwEIWjVu+tZOXOCrQKGY8uSGdGqhm9So5BLaPbNuDQMyP11HZYuWZcLGEGJV1WF3vK2vjJzBS2FTWxtaiVBp2dXqcHi8ONXjk0v1wiCCSEavjgcC12l5eMSD1GjYJ3Dtay7ngDl2SEMystjNd2VtBtdzM/O4LS5l5CNHLuW3OY31yawcUZ4Xi8XqxON6/vqfKP/fK2Ml68bhThBiVN3Q4+OSYmR2USgXHxIQRr5MSGalBIJcxMNbOtuAWpROCmiXFMSzGx9ujgJgAahVgteqCyI0DH2+eDXWVtvHTDGACigzXUdVrpVSvYW97O9RNisToGJ6eO1XZhcXjOypGPCNcjEQaKmmammVmzrzpgH4/XR1Fjz7Bw5BfwfwPnzFrx+XwHfT6fzufzzez7N97n8606D7b5NYh/COwua/M7cYD6Tjur91YNqaGRPnIs+yoGl+5WtX17zvez4/W8tLUMu8uLD6jpsLK1sJk9ZW388YocRscGo5JLmJ8dwc2TEsiKCmJPeSsvbSllQ34joTol3VYHG/KbCO1jY9w/ZwTbi1tYODISjUKG9qQqQ0GAR+alUd5qYX1eI09+lMedbxzkcE0nr2wro6rNysqdFfz+0wKWjo0hPULPG3sr+eJEE4kmLVNTQjlS08X7B2tp7HKw+UTzEL+pgQfmjMDc9zaiU8r44xU5aJVSbpgQx/Mbi/nVR/lcMz6Gfy8fx5+uyiHcoKKy3cqc9LCAkJQgiOGi9Eg9CtngUMTJjj3JrOUX8zOo6Cvx31rYwui4wc50UU6knz1zpsiKMvDa8vGkhGnRKqSkhuswDqE3rj2HsNr5xA95D50JLth5bhgeV9vXoD8Z80Mgr65z0LavKtqxu9xoT1npyn0OpqSY+DwvsDVW4tfIvvajx+7irX1iyyqFVMK9s1P4y4YirE5xFTkz1czsjDBumBDHopGRbCtu5vmPi/3NiZt7HDz5UT5JZi0SQeD5TSI/PDNSz/PXjMLm9GAyK/jbdaPZUdJCW6+Da8bH8uq2cvZVtpMTHcSvFmfylw1FbCpo4v45I/jbZrFNVlOPA61SjHFPH2FmaoqJ1XurmJJs4tUdYkiksLGH6akmdpW1BfyuhFANggC/WpTRl7T0UNDQxardAyv3/n6Zd//3sH9bilnLpSOjuGtmMq29DmxOD0kmLTanmyc/KuDxRRl8eqwhILQyK83Mvop2xsQFI5NKmJpi4sO7p/bxv2V8cLCOmybF8/b+GpweL9NSQrlsVNRZ66DIpBJmpoURqXBiCAnFrFcSrFFw5xsH/fskm7XkRAed1fjnGz/kPXQmuGDnuWFYi2b1sz9+CExKMg3atiA7YpATB9iyaSMPXDyCRJOYiRYEuHVaop+v/XVQyCQkhorHXJIZztv7a/xOHGBbcQs+H7y0rRSnx0uQWhHQYR5Etkd9p53/fjXwel/Q0MMTa/NIMmvJjQnmg0O1bC1qJjVcz+Mf5rGnvA2P18eRmk7+sbWU6yfEcbyui1CdgrTwARpUkFrOguxw7C4PG/MbuWdWCluLBlbgnTYnExNDCdEMnJNQrYJEk44thc3ct+YIr+2qIDJYRaJZS060gZgQMc8wJSWU/5wUkgGYkRbGyp0V/GVDER8ermNLYTN/XF+IVCLhnzeNJS5EzYqbxnLpyEjmZobz+MIMXt1RznUr9nC4euDBa9YrSQnT0WFx8/T6QvaUt3HbRYncNycFqURCddu3d6UfCm29Dsqaeylv7qWyoQWjTiE69lQz7/1kMr+9LIsXrxvFv5dPINY4PFTyfsh76Exwwc5zw7Bekf+QmJAYwo+nJvD67kq8PpidZubKMTFfu396hIF37pxMVZsVjUJGklmDSv7Np1cpk3LXzGS+LG1l2ggTn+cPbnbr9HiZnRaGXiUnSCVDJZcEJC8FAXTKwbHeLpsLqUTgREMP6/reFNxeH7UdgSJdrb1ONAopk5JCaei0MScjjKKmHsYnhFDdZuHXnxT49/V4fcQEqyhtFqlcGrmMPeVt3DAxHplEQBDEpstv7avmkflpTE42oZJLaeiyIxUEIoPUpEcaSDHriA5RsTG/iclJoehVorBVmF6FXiXD6xMZKvVddt7eX4NMKiCXCpQ0WWjosDI/O4JWi5P6DltfcwkxgTs+Uay4s7s8dNrEghuJIFDa3Ou3GWDJqG9uBHwqfD4fe8rbeOz941S1W8mNCeLS3CieXJvPfXNSiA7RMC7ByLiE4V+ZeAH/f2JYO/KsrKwf7LvNehWPLkjn+glxuDxe4o0avxb3qRg3eTpFjT0EaWRnfDOPiTfyyT3TOFzdwZTkUL4saQ34PC1cz+i4YCwON2/tr+bXl2ZS2myhy+Zi/fEG7pqZTFSwOoCnDBBnVBMTomFP+cB4EkEISNT1o79pREakAbvLw2vLxxNpULLo7zv9+xjUMmanh2E2qNhT3o7D7aWx206yWcvvPj0RMN6TizNwe71Ut1v55Gg910+I4+9bBpgjeqWM/9w6AYkg4Z87ygnRyFk+JRGr080be6r89L7MSAP3zkqhstXCi1tKGRMXxKz0cO5bcwSP14dBLeOBOak8u7EIdx+PuKC+i+c3FbOvop2pKSaeuyaXh9496qc6xhrVQ1Z2fhPKWyzcsmq//wF6tLaLXoeHUbFBrDveyO3Tv72j/A+FH/IeOhNcsPPcMKxbvXV1dREUNDxijV+HvLouHn73CCcaezHrlTy5OBOdSkqSSUd86DfHyPvR0mNn4d92cuu0RD473sDxui7Ucin3zUlhd1krGZFBTEoMweuDL04089GRekJ1Cn52cSpxRjUjwnUcrO7koXeO0mF1EWtU8+AlaWRHGbC5vVz+0k68PjFRd9EIE9VtVuKMGryIPG2L083+yg4iDErW5zUBcO/sFLpsTv6zRwzZ/GJ+Gi9tKUWjlHHz5HhcHi9hehUSwAus2l2J1eFmyehoxsaHUNLUyzOfF3LNuFi+LGkZ1NTi8YUZPL3+hP+hkh1tICFUy6fHGgbtV9DQRWSQmqhgNU+szQv4PCZEzUUjTCwZFU1cqIYrX94d8F1JZi33zU5h7eF64k1aTFoF87IjSA0//Uq6LSeauOX1wdfrvbNT2FLYzPt3TTmvzSDOJ/433ENwwc7TxdfxyId1jLy/InK4oq3Xwf1rDnOiUXxtb+lx8PN3jnC0posrXt5NQX3Xt47hcHkoqO9GpZDw5w2FxISouXd2Cj+aEo9KJiEzMogVO8rx+GDTiWbW7K/B5vJQ22HjwXeP0mlz8d7BOg5WdvDIvHSevyaXJaOieezDYzz47hEUUoHfL8nGrFMyIdGIIAiUtvRS02EjTK+kus1KqFbJFaOi2VjQTG5MEA/OTcWHmCeYlxWOSaegrtOOxemhpcfBcxuLeXFzKZ/nNfLB4Tpaex2kR+iZkmKi0+riH1tL/TrrkiGEqGAwTz7OqKGgfrBmTXWHlS6ri1e3l9F9inYLQG2HjcU5kYyOC6Gy1TrogVHeYkElF7sOBamkzEoPOyMnDhCkGfwmplFIcXm8jIkLRvE9ytKeKYb7PdSPC3aeG4Z1aOWHRH2njT3lbRyp7mBcgpFJiUbCgwILghq77JS1BFIMXR4fbq+PdouTN7+q5ndLsr+xgu9EQzf3vHmYh+al8euP81nfF882ahUsmxSHpk8m1eH2snGIGPrx2i5e3T7QZejHUxOQCHD/7FQsTjcPvnOUuk4bD1ycyr7KNj47Jo5R3NTLnvI2rhkXw9HaTjosTuKMGqYkm3iuTx0R4MdT4pk8O4Wihp5B391jF3tsdlicbMhv5ObJCWRG6gkzKEk263hzXzWbCpq4dnwsL28r8x+nUUgJ0chJMuvIjgqipLmHozVdLMyOoHxnRcB3hOuV/HdvFV7f0A0pYo1qUiMMKGQStEPkCgQB4kO1zM+O/No5+Dakhum5aVIcb+wdSCjfOi2RrYXN/OXq3GHRCegC/m9jWDvyH0rWssvq5IkP89jSx9B4Y281V42J5qkl2Wj7JEp7bC6Ucgl6pYyeU1aX/Su0Y3VduDxeFN9QeFLZZqXH4cbr9fHYgnRqO2zolFLUChmvbCvjFwvScHt9yCQC4QYVHdbAValEIjYc7sfqvVXcOSOZZz4v5NmrRzIuPoS5WRFoFVLWHQ98ELRbnChlUj7Pa+TBuWnYXB7+Z0dgteW/d1fx6IJ0ksyDw0RzMsJIC9ejVsiYkWrmiY/y+FefI040afj14kye/DifwsYeHpqbxu6yVkw6JdNGmIgJVpEZaeBAVTtZUUFcPyGUJJOWBJOGxm4HcqkEo1ZOm2Xg927Ib+TO6Un8a2cFbq+PYI2c567O9fPVI4NUXDMuhndOqrK8dWrit9JAvw16tZwH56axICeS+k4bRq0Ch9XC1cvGEjdM2Clfh/8N0rBwwc5zxbB25EuXLv1Bvre0xeJ34v14/1Adt0xNJClMy86SNv66qQiVXMpjC9N5fG2en9t85Zho9pS3+f//TU4cBhKNX5a0sDg3itoOG26vmJgM0cjRKmTMSDXhdHu5cWIcv/mkwF+UlGzWYnF4AopiXB4faeE6FmRH8EVBE3FGDc9vKubumcnIJZKA/qAgrlhDtUrSI/ScaOj299c8Gb0ON1sLm3lsQTqfHKvH7vKybGIccqmEJ9bmMyExBLNeSX2nnQiDirlZ4Xi9PjptTn65IB21QorH4wPE1nFj4oP547pC8vpCKbUdNoqbunliUSa//6zQ32FoTnoYI8J1PDwvjT99XsTxui5cHg9v3zkJl8dHdLAaj9fLpr43laO1nVgcHh6dn4YPkUmUGxt0XuLXwRoFU5IHU1KHO36oe+hMccHOc8OwduSrV69m2bJl3/v3ntoHsh8uj5fDVZ3c/p+BxJcgwIvXjaK+006IVsG6Yw3sr2jn+vGxzMsMbHZsd7kpa7FgdXiICVETGawmM0rPNeNiSIvQ8+C7R/1O2aCS8eelI3l6XSGPzEulrNnC2IRg/n7dKE409iCTSogLUfPHdYGMkUlJRj48XE+XzcWt0xK4760jAGwsaOKacTGsPolvnhImys7eOi0RQRC7FXXbXHxxotnfJEKnlDEz1UycUUOvzc2j89IIUiv45FgDu8paael10G5x0djtYH52BDEhat47WItMInDz5ATGJYTw0pZSMqODGJdgRC6REB2k9jvxfmRFBfHshqKANnGbC5uZnmpGEOCJRRm8trOCJaNjyI0Ri38OV3dw07/2+ePtY+KCyYg08MznRWgUUj64ewpG7emJEXVYnXT1tWgznEbV5w91bZ4pLth5fjFc7RzWjrxf8vH7RqJZy4gwHSUncY/HxoeQYNLy0pZAASaDSk5Tl4PWXic7ilsYGRNMVnQQbb12Ik8S2eq2uXhjbyUCAk09DpxuD/OyIpiSbOKReWn8ZPWhgJV1t11kksQa1dR02Fi1t4qXd5Rx8+QEVvSFP8bFh/DUkmw+PlLP8bouJieFEh6k8tu4bGKcn5JY2txLslnHH5Zkc7imk2SzlphgNdUdNjYXNhEXquF3n52gvLmXWy9KxKxXUtbSS060gW67i1+tzUOvkvHTWSl8eKiYilYLF6WauCQznKLGHnKiDdjd3gCBquc3FROiyeKmSfHcu+aI3+H++tJMBCGwvD7OqGHd8UDGSv95e25TMVqFlEfmp7N4ZAQyqQSr082zG4sCkqaHqjuZ0ef4rU4PVa1W0ofQxjkVByrbeXxtHkWNPYyKDeZ3l2eR8y06KT/UtXmmuGDn+cVwtXNYO/LvEg6Xh2O1XeyvbCdEI2d8opGUMJHNEKZX8fKNY1izv4YvS1qYkxHO1WNjCNYoCNWJjQbmZUUwLj6ETpuTf2wrxeH2cuXoaHrsLnaUtHLbtERAXMUXNvaQV9dFmF7Fvsp2v1Lemv01rLl9IsEaBXOzwlHJpewsHeB9210efjIjmR/9ez8er48bJsSRGxOMUiZhTkYYOqWce948zILsCG6eFM9b+2uoOKm7/dHaTiYnGdlTLurAbMhvZGxcMPsr21l/vAFLXxXpYwvSefjdo9jdXh5dkM7HR+s5VtuFWackdn4aWwubcbi9/GRiPH/6vNDPp153vBGrw0NcqNgZZ+sQgle7y9rQqwLbyX12rIFrx8WyZn+Nf5vT42XaCDM7ilsCju+Pf1ucHn79cT5ZUQZCdSp67W5ODJGA7ba7UculWJ2eIVu0nYrqNgu3vL7fL052pKaTO984yNqfTvUrUV7ABQx3DGtHvnz58u9s7C9LW7ntJG6wWa9kze2TSA7TAaLK3eMLM7C5RqBRyPzMk9kZYWgUEt78qpq0CD3Pbyrxj7H6q2oempvKjFQzMUY1Vqeb3aVt3PHGAXKigzDrlUgQWDIqmu3FzfxyQQZbi1p4bWclTo+Xi0aYuGtGMq9sL0MQYGKikV6Hy+/Ej9d10W5x8MK1o2jstvPbvqrLrYXNxISoA5w4gNcLP5qSQFSwmq8qRG0Vg1rO0rExvNLHIokOVqNXybA4PczPjuCjI3Xk1Ylhj5ZeB794/xiPzE9nXV4jXp9vkCTutuIW7p2dwn/2VJERaQA6Aj6PNWr4siTQOR+o6uDa8bFMSDByvL6LJJOWgvpupiSH0m5xkFfXjVou5Z5Zybi8XhTSgdh+RauFcQlGjFoFi3IiApgkIDaRsDo93DgxbkilylNR3W4NUJgEqO+yi/TMb3Dk3+W1eT5xwc7zi+Fq5/AlwAI7duz4Tsbttrn4yykty1p6HOwqbeVEQ7e/c41EIqBVygPog4LPx9PrC7G7vRQ1Dl4Rrs9rpN3q4LbXD7D5RDMvbi7h8UUZmPRK6jptxJu0zEwzce/sFJp7HAHUwS9LWumyO5mfFcbjCzP42+ZSgjUKxsWHkBapp8PqJMGk5alPCyhuGvhui9ODw+1NstM7AAAgAElEQVRlbFwISpmEH01J4JH5acxMM7M+r5HD1Z2MizcyLyscq9NNcWM3f7tuFPfMTmFGqpnWXrGSMiVM53fi/fD6BroODdWyTK+UYXN5KGnuJScmKEBpMVgjJypIxYiwwbztpm4Hv/4kn2vHxXDVmBimpITy6rYy4owaHpybyt+uG8UnR+vZW9bO44vS+dnFI5g+wuTXd5dJJdwyLZHZaWZA7D706IJ0sqL0rLl9Ig/PSyPoNGLdQ8XD5VIBg+qb1zjf1bV5vnHBzvOL4WrnsF6R97c/O99wur10DlVc0mnjT6/u5o9X5HBZbhSCINDcY+er8naq2nqJD9VR32nDpFNy3fjYQatTgHijhqyoIFLD9OTVdXHbRYk8+sFxvxjWiYYefjw1AYVUoMM6WEltZ0kbS8dE88z6Qu6bk4LL4yPOqKG+w8bTV+Tw3MZC2i1OIoICV4v/3lXJLxak88AlI3h+YxELR0axLq8RtVzKT2Ym89KWUqanmjhQ1YngA5VcglIqYe2ROhaPjGJktIEeuxuTTuF37P0IUsu5fFQU1e1WxsQFc+gkgaofTU3gvb5Q0UtbSrlntth9qM3ixOvz8ecNRdw6LZGpyaHsKmtDJZewbFI8u8ta6bG7eWlrGS9cO4rFI6Owu71UtVlptzh56N2jdNvdTE8L4/lNJXTZXCzIjiD+JLpfoknH328cQ12HDaVMQpxRc8Zdd1LCdNw1I4lXTmpy8ci8dBK+hbL4XV2bg+D1QGcVeNwQHAfyMwv3fG92niMu2HluGNaO/LuCSa/klqmJPL1+YFUulQgEq+VYHB5++cFxcmOCiQwWE4frjjXw+OIMHl97nHtnpXDTpHj+sqGIe2enEKZX+hUJNQopczLCyK/vpq7Tht3pIdGsDVA0BHjzq2oenpeGXBb4MEk267hqTDQTEoIZHT8ej9fHrasOcN+cEVS1WfjrFyVcnBlBVrQoQjUhwci+SjH+HaZXkhmh57WdFSwaGcXT6wv9NMV3DtTw1OXZfFnS6k8oTh1hwqCS8ezVudhdHi7NjaS4qYefX5LKE2vz/MU3i0dGEhOiRimTYNarmJBo5OKMcFweLxmRBj45Uk9jt1hN2WVzEW5Q4fb4eOGLgZDT37eUsiA7nBU3jeVQdQfrjjdS3S4mjcpbLDg9XjRSGdlRQfx5fREtveL5TAzVoJBJ/D071+c1smR0NElmnX9srUJ2xpWaJ0OjkHHXzGRmpIXR2GUnJkRNRqQB+XCo1rS2wf7X4MtnweOAnGth1uMQEvdDW3YBwwzD2pHPmzfvOxv7itHRyKUSVu2uJFSrYHFupF9W1eL00Nhtw+fzUdzYw28uzaDL7uHK0TEkmXU8/N4x3F4fL28r45apiSjlEsL0SqKD1by9v4btxS0kmLQsHRuDfIhwhEouJT1CT6fVSW5MEHn13fzs4lRKW3p4/1Adzd12lk2K5w/rTnDzlHhW763yO8tD1R1clhvJJRnhuDw+LskMR6+WEROspqS5hw6rkyM1nQENMLw+2Fve5qcUArg9Pqo7rOjVcipaLXxwqBaJINDW6+Bnl6TidHuRSyVEGJS09ToxaZVsOdFEc4+du2amcKy2k21FzdwzewTT08TwTFaUAYlE7Gt5clwbINmsRyIIfHGi2e/ExXmIQtNXZJURaeD9u6eQV92GXKHgQFW7P5bfj9LmXuadg25RQ5cNp9tLRJDK3yHIoFYwKSn0jMb5Lq9NP6q/gq2/H/j72BowpcL0B097iO/FzvOAC3aeG4a1I+/vnn02cLm9lLX2YnW4iegTXDoZYQYVt0xLZFqKiWc3FvL0ukI/VW9sXDDN3Q5WflmBIAh4EbC7PDR128mr66K9T53P4fbyynbR0bx0w2j+58tydpaKxUBVbVbKW3uZlxnB8ikJvLWv2l9sIzZOcPLoB8e4NDeKW6cl8sLmEsr7yv3LWno5XNPJk4uzKGvp5upxMfgAuURg3fFGPj3WQIJJG0D1e3RBGsFqObkxQTR0BWqWg0jH63+omPVKMiMNpEfqee9gLeUtFiYlh5IQquGxD46T31eiH6yR89drRvFFQRPhQSqmjTCRZNZysKqTz/Ob0CtlbD7RzPRUE9OSQ/nl2jy2FrUQplfy2MJ0dhS3UNdpY3Z6GLUdNv6xrZRnrszhhS+Kabe4WDYpjvnZgVz7OKOG7tZGXEotTV32QR2Z0s5y9W11ull/vJHff1ZAt93NlaOjuXfOiLOuzDyXa/O0Ufnl4G3H34YJt4Pq2xO58D3ZeR5wwc5zw7B25Lt37yY7O/uMj+u1udha3MJfNxVT0WZhSlIoP78klbEnScy29NjJr++mzeLkuglxtFucHKjqJDJIdPD3vHXYz3PeU97GowvSEYDc2GCSTNqA1W3/orvfiZt0Cu6amczKLyt4Y081ySYtz12TK67UQ7XkRBl4fU8VdpeXdw/UEhWkprzFQkyImvQIPaXNvRyt7WJ3eSsjo4P41UcFONxeJALcMT0JmSSQgw3w0eF6HrgkldFxRhKtTjYWNAV8vjgnkg+O1LJsYjyTk0P9jJRuuxinP1DVwb2zUnh0QTp2lxeVTIJJr+T9QzWMizdS3W7ltV2VAMzPjmDxyAhyY0N4bWcFb+6rZnxCCJdkRrCjpJXmHgdPfVrAn67MYXNhC2v219DZJy3w4uZS/rI0F6/Px9i4YLSqQElZi8PNmuOdrN5/gl/MTyPZrKOspRdBgJsnxzMq9uz6YB6v7eLBd4/6/373YC3BGjmPLcg4K62Us702A9DTBI5u0EeAcogHlDlt8LbIUSA//YfPebHze8AFO88Nw9qRny3yG7p56N2j/hXwrrI2rM4T/POmsYQZVLT1OvjlB3lsOjHg7H53eRYz08JxuNxsK2pmflYEyWE6Spp6+eJEE1sKm7l2fAx3rj7IU5dl8/K2UqrarBjUMu6cnkxFiwWdUuRLXzs+juc2Fvtj42WtFn73aQF3TE9CQODDw3X+UAmID4J7Z6fQ0uPgSE0n4xKMXDEmBq/Xx9HaLv/v8PrgnzvK+eeysUglAn+4IpunPhGd/NJxMTz87lFsLg93zUjm+Wty+fBwHT6fjyvHxBCkkaGRy8iI1PObT/K5cWKc34n347VdFTw8Lw2VTML7h2op6ONpf3qskavHxjA6NpjDNZ18ntfIP24Yzf1rjvjfYvZXdmB3ebkkM5zP8xrx+aCmw8aGU4S+mnvs7CptI0SrYHrq4MbUpS29rN4vxvGf3VjMlaOjWTwykqkpoeREB6FWnN0lm98wWFnx/UN13H5R0vfPF/d6oHQzfHo/dNdD/FRY+BcI74sZdddD8wkwJovbmvLF7eoQmPRTkP5/edtewDlgWF8Rubm5Z3VcZZtlkGbI4ZpOavu4wTXtVpp6Ah1pbYeNOKMGl1dCaoSBV7aVsT6vkexoA08syqCm3UJMsIbrx8fyz+2l3D4ticYeO3aXh1W7KpFLBe6emcyfNxT5KwtPRlO3gzijhle2laKUSbl7ZjL59d3sKG5BLpWwpaiRozWi7G1hYw/JZi1XjI7xN0zoxz2zUlh/vIF1eY3EhGj4wxXZ/HVTCQqZxO+YX9xSSnSQmscWprOrtJU/rS9kXnYEC7Ij6LC66LAEslImJhqZkhyKIAiMCNdzoLLd78T78f6hWu6ZncLhGpGx0mZxBjSyADhe18WMPjogQIRBNaiRxaUjoyho6OKJRZmUt/QSolUQohlYlXedJArm8fp496DIiJmZZj5rJw4Qphtcqp9k0qIZQjHxdHC21yYgOuk114O370FatQs+uR9ufB8szbDmBmgtFvUfpj4AMx4DiRTM6RB6Zk0szsnO7xEX7Dw3DGtHnpKSclbHnewY+qFXytAoJLx/sJbVe6sw65X8anEGr24r56qx0WwqaKKsxcKt0xL5++Yyf5eavLpuWnrKuGtGMre+foCMSD2PLcyg1+5G5xRP302T49lW1EyEQckj89KIPiUeD6CUSZBLJfxifgZ7Klq5+7+HEASBxTmRmPUKvxPvR1mLBR++AO72+IQQSpp70Cpk3HZREvWdNp78KJ/nrs6lsi2wGOi6ibH87J0j/s44/9lTRbfNjc3l4kdTEnB5fIRqFUxODkUA/trHMkkya7l7RvIg+70+EBBtSTZrhzzHwRo5SpkEmURgQU4EPuCxhRm8+VU1DV025mZGcNmoKFQyCfe+dZjCxh5Sw3U8feVIxsaHACJ906CSBbwtxBrV/l6fZ4tRccHkRAdxvE48z0qZhEfmp6Ebogfr6eBsr00A2ssGnHg/avdDTz0UfS46cRDjZzv/CtMfhtlPnNl3tJVC0wnG6gTorIHg2LO393vAOZ3P7xHD1c5h7cjff/997rjjjjM+LifawCWZYWwqGFAwfP7aXCrbrPzp80I/XXBnSSv3XzyCHpvbryuuUUj9TrwfTd0O2ixOeh2i/kl8aBMmnYJX+7jHggBPXyHGg4/VdRKqUfCT6Un+bvMAD81N4+Wtpeyv6mBGqpkHLk7l+U3FfHS0nksyw4f8HXFGDT4gRCMn2azjtmlJ1HfZeGNPFeWtdSSbdTx1eRZtvQ56HR5Sw/W09jpotzhxebx+J96PT4/V8/KNY5BJBA5UdXD7RUlolVJ+9VG+f5/yFgt2l5eYEHVAf8/5WeGUNPewZFQU6REGthU2sygngs/6pHEFAX59aRaz08xcMToa8HHJ81/iw8f8rEguzghnZ2kLvXY3D32S7+eqFzf1cvt/DvDJPVOJDtEQb9Lyp0UJvLCriaLGHsbEBfPU5dmY9ecW/ogJ0bDiprEUNHRjdXoYEaYjLeLsaYtne20CoBmCIaMxglwrhk+mPwRSJZRsgNoDULlLdOqny5GvPwL/uRzsneINHpoC178Npu/JCdl7xPCP/PQfvud0Ps83OqpE7r4qWGQJncTdH1Z2noRh7cjPFpHBGp66LIurxsTQbXMTGaxi9d4qSpp6WZATicfrY/XeKhxur79Rbz/6mwifnEyclGgkNVzHzy9JxahVIJcK/OL94/7PfT74/Wcn+M1lGWgVUkbGBqNVSHhwbip2l5eMCD2vbC8jv0/xb1tRCx6vjytGR1HabOFYbSdzM8MDEpSTEo2sz2tkT1krK380nn/vqqDb7uJvm0votLqIMKi4bkIse8raMOmUjI0Pwen2IJNISDBpkEkG86BDNAp2lLRytKaTnOggooPV/k4+J+OlraU8d00umwrEcM+0ESYUMoHGLgfjE0LYVtRKVpSeWKOGORnhSCUCdR023jtQQ2SQilW7KpBLJSyfEs+a/TWUNPewsaARqSBgc3kGFRy1W5xUt1uJDhGTePXHdrLmjpvOSI3wdBAZrA4QMvvBEJYFuTfA0TfFvwUBFj4PdQfgswdEZz7yWhh7CyRfDPrI03fiHjd89SrYB4q2aCuFsq3fvSPvbYHCT8Xv10fARQ9B/BQxLNQPp1V08GdYuPW9oXovvHUd2DpEG2c8BpPuOm2W0A+F8+LIBUGIAH4P5Pp8vvHnY0yAsLDBybDTRWSwhshgDaXNPVz+0i6/QFR5q4Ulo6LJjDRQ0NBNhEFNsHogTLC1qCVA0Gn5lASau+3ct+YI981O4c8bCrl5csKg7+t1uNEp5SSF6fj4SB0/n5vKuuONbC5s5qezUvxOvB87S1t55socOqwuEs06QnVKsqODqG63MCJMT7vFyT93lLMwO4L2XgcLsyPRqmQYVHI6rS5umZbInz8Xi34eX5TBT1Yf9Mei1XIp/3PzWNLC9RSdVMp/85R4/rWzgk6ri4szwnh1exk/uyR10G/Jjjaw8ssy7pqZQllLL063F6lEikmnxOr0sCAnnP/srmZXmSjwlRCqYenYWHaVtREdoqGgoYeaDit/ujKH5VMSyK/v5orR0WLMXCIQZ1RT3T7wAJEIot53P0JDQwnRKIYM3wwXnMu1iSYE5v0Bcq8DSyuEJouslVengSEaJtwJu14A6ysQnABL/nH6Y3sc0Hhs8PbWwsHbzjfyP4D1j4j/bykU6ZO3boLosdBWLvLgCz+DxOkw5kcQlu4/9JzO5/mCpQ0+vk904iCu0Lb9UbQ3fjIwTOwcAudrRT4N+AgYdZ7GA2DJkiXnPEZhY4/fiffj02P13DE9ibpOG6Nig1izv5pbpyXy5lfVHKvtZHqqiX/eNJa2XgdN3XZW7e5nXgh029xIBWFQwUtMiJq95e2s3lvFopGRtPQ46ba7ee7q3CHlAEw6JcfrutlW1MK2ohZ+fWkmnTYnJ+p72FXaxtIxMfx56UgkwL92VbKvoh21XMpNk+Op67BS0tyDy+NjYqKRL040ByQUbS4Pe8vb+fXiTKo6LJS1WDBqFHye3+inAXr7WCWRBiVLRkWz9kgdMonA0nExzE4L482vqiht7mVUbAgfH62nqk0s4okKUvHUkmy/Ewexy1G/TG5xUw9xRg3TUkys+LKC0j4p4I0FTSzKiaSlx87PLknloXeP+TniP7s4laSTSuLPx7x/1zhnGzVGSJoBjflQsU1chbusMOpe2PI78PS9tXRWwgd3wO1bxFXut0GhhdzrYeMpMfXk2edm77fB0gp7Xgrc5vVAzT4IHQGf/RzKt4rbm/KgeCP8+DP/bxoWc25rh9aiwdu7BjpODQs7h8B5qUP2+XzvAeedKb9q1apzHsM7RKNHtVxKdnQQa+6YRGZUEDdPiic2RMXdM5P57WXZjE8I4W+biqltt7G5sHnQ8W/vr+GheWl+idVks5ZbpyWyZn81bq+Pj47UU9DQzcWZYdR0WMmNCWJaykB3GUGAW6Ym8MlJsq/vHqihvMVCq8XBpKRQgtRyNhxvYHNhM/sqxDJ8m8vDih3lTEwM9TtBlVyK7ZQHFYjhil+uPc7GgiYqWiz8eUMRx2rFRF+EQYXX5yPOqGZEuIErx0Tyi/lpvHj9KE7Ud3P3fw/h9vqIM4odiPqdOIjKgF+WtAxKPubXd3FpbiQTk4wUNHQTZlD6nXg/1uU1MCk5lDf3VrP27im8eN0o3r1zMj+emojypC4+52Pev2ucFxub8mHVAtjwS6g/LDo1n2fAifeju06kJJ4uMpfA+NtBIhM55xf/FuImn7u93wSJHBRD5BzkGmgvH3Di/WgvhZaB3rDDYs41JggfgiMePCCJMCzsHALfW4xcEIQ7gDsAIiIiWLFiBQATJkzAbDbz2WefARAXF8fcuXNZuXIlra2trFq1iuXLl7N27Vqam0WnetVVV1FZ30xtL1g8EpLC9ESqvWxYvw6ApKQkpk+fzpavjnKiV0d8qCbAGd03JxmJz0dZfTvVLV28daCBbruLy3KjKGvpoanHTlmrhYxIG3Mzwv2KgB6fjxCNnMZuOy9tLWHJqGhGxQbh88EvP8wLoDx+crSe8QlG9pa3IREErh8fw6KcSDptTsINKl7eVubXEAFxhTwx0UhUsJpNBU1kRhqINmpYe7hu0LnstruYnRbG2sN1fFXRxt0zUzhSMxATFRsOa3hzn5XKNis/mZFEklnLtqIWRsYEMSs9jD1lbfxibiqH9+9lY6ueudlRPPTuMb9NO0vbaOx2cNWYmEHfn1/XPSgZOjY+hBMNXVwxOoYPDtYNKliCvnwdAo09dqoLj6DBTW7mdFateg0AjUbDsmXLaGpq8l8f1157LXl5eeTniwnZGTNmIJPJ2Lx5MwCpqamMHz+e//73vwAYDAauu+461qxZQ3e3OG833ngj+/fvp7i4GKlUyqxZs3C5XGzfvh2ArKwssrOz/R3SjUYjS5cuZfXq1f5GAsuXL2fHjh1+0SSHw0FeXh67d+8GRFpaSkoK77//PiC+gi9ZsoRVq1bhdIqO+bbbbmPjxo1UV1cjl8u5Kc2GzN7HVjryJsx+Ap/XxaDosVJPQWUju9ev5LbbbuPTTz+lvl507Jdffjm1tbUcPHgQGLif9tknknnxbGwOFyMnL2blv/4FgEKh8N9PLS0tJEabmTY6nYZOG1t2H8Lj8TBlyhT0ej0bNmwABu6nfifWP0/vvfce7e3t/nmST34QzUe3DNitMVLmMiOpriHx1MQTgFTmn6fW1lYsFot/ngDmzJmD2+0+p3maN28ePT09pzVPbrebSyc+TviWB6C3CSQyeic9hF0WQe2RI+zbt4/W1lYOHjxITEwMH330EQBRUVEsXryYlStX4vV6kUgkpz1PQ/m9U+fpZL/3dRB8Q91xZwFBEGYCz/p8vnHftu+4ceN8Bw4c+LbdWLFixZAZ4k6rk99/doL3+jjGggAvXDOKy0dHB+y3o7iFH6/az53Tk/D6fLT1OpmUZOSjI/XsKGnl0QXp/PnzwoCwxM8uSUWnkKKSS/n4aJ2/V+M/t5fRYXXx6IJ0ajusVLRamJcVwdGaDoK1ykGdgyYlGblqTAyHqzup77IxJyMMlVRCXKiW1l4HP3/naIDj/+1lWVS09rJqt6j38rOLR3CkposumzNAbRDgT1fmIJEIONwetha2kGDSkBCq5e0DNWgUMq4aE82/dlZQ3DSwIp6ZaubBuan8YV0Be8sHNMNfvG4Ubo+XyjYrL57yGwD+fv0o7u1rF9eP31yaSUWrhdf7tGlyY4KYnmrm71tKUcokvHDtKLrtLl7eWkrVSbHwORlhWJ0eZqeZuX36YIpjP75u3s8ZzSfEZJazF2InQdTosy6uOS827ngOtjw18LdcA3OeBFsnbH9G3CZI4Mr/EYuGehpBZ4agwQ/Xs7Kz8Th8/pgYyzZnwuLnxOTk2cJlFxO2ZVtBaxZjy+GZYoJz/S/g8H8G9k2YDlevAm3ot9v5faOrDjqrQRUkhoVkA8n2H9pOQRAODuVjh7Uj73+6nYrdpa3csPKrgG0hGjnv3DmZEX1aHN02Fweq2jlU3YlcIvDJ0QbGxgeTX99NfZcdpUxgTnp4QA9LgIkJRi4fHcUvP8zzb5NJBF66YTQlzb0IQKJJy76KduKMGspaepmUHMoLXwxopajkEp69OpfffpxPy0kMjVunJjI+MRiTVkmLxcm6Yw302N0szIlgRJiOH/17v58/PSstDAQfl+VG8cSHef44/4REI2nhehbnRGB1e6lpsyKXSTBqFJS39gI+BASe+Tww1nfpyEgyowz86ZTtUUEqVt82kX0V7Tz6wfGAz+RSgVeXjaXL5qKq3cpXZW3EmzTcPyeVEI2C/ZVtHKzqpKyll8+ON/gXXC9cO4olo6MpberhwyN17C1vZ3yCyBM3qORcNTaG8G+opvy6eT8nNBXA64vAKq4gkUhh2YdinPoscF5srNkHr80D30lFX5f9HbKuEpOFvY3ia729G979EVhaRAd5xQpIOb2Y99faaWmFfy8MjAkr9XDHNpGueL7R3QCVO8WHRvRYSJoVoOL4ncz5d4Af2s6vc+Tni7UyA7gJiBQE4QngOZ/PN5jXdobYuHEj8+fPH7S96aTy9n50WF209NiRSgSO1naKgks+KGroJjc2mDtmJBFhUJJo1lLTLmqKxxrV/HdfdcAb382T43hhc+DK1O31caCyg7RIPelhemq6rKRHGghWi511ylss3DMrhaYuGyMiDHRYnLRbnCybFM/re6r8Iltv7a8mxqiix+7mk6MNmPVKuuwuHvswj1/1NZ/od+Rbi5pZOjYGg0rGj6clIhFAKgiUtVh4Y28Vs9LMPPjuUTr6kpeZkQYenpfGo+8f47lrcv0aJQBmnZKsqCCsrsH6571ONxqFjEsywzlU3cE7BwYSO/fOHsFXFW2s2FGBIMCSUdHcMyvFL0AmEQRW7a4MCBGByMUHSAnX8/C8dDxeH712Fw6397TK4b9u3s8JlV8OOHEQE3E7noWYCaA4c0riWdno6BH54bI+Nk7UaLjpQ3Flbm0Vy+/TFoBSCzFjxX3aK2DVIugPwVha4N2b4c4vwZh49nZ2Vg9O7Dl6xO/7Lhy5IRJGXi3+OxM7hxmGq53nxZH7fL7twPbzMdbJqK6uHnJ7dIgamUQIKBHPiTbgA65fsZemvoIfs07Bn5fmcqSmk9YeByqZhNoOG+EGFU6PF6/XxyNzU1m5s5KbJsXj9HhRKaRDalEb1HI6LE4eW3ucmBANV46O5tEPjvuddGSQil8uTOeZ9YX+JJ9aLuXheWn87rMCfD7w+nxEGNRYnG56HW5cHi+LciLpsbupbLPy5KJMnvw43y/zuqe0lflZ4fzPjvKAMExuTBDvHqz1O3GAgoZu8uu7+MvVI/ndpwVMTg7lzhmJ1HWIMgJ//aKYh+elDWLbXDM2FqVMIESr5BcL0lk8MorGbjuxIWpaex3+sIrPBx8ermN8Qgg3TBQZJjEhGn48NSFAezw7ysDImKCAcyeVCASdAZXw6+b9nHCyE++HpakvsXjmjvyMbOxpghMfwYHXRDrhtAcgdiJI5ZA0E2Imgs89tHBWd92AE++Ho0dkUpyGI/9aO5V6kKnAfcqiSH12omTniu9kzr8DDFc7/1cWBGVFGfjdkixe+KKEpm4Ho+OCWTomhh3FrX4nDvCTmcn89M1Dft2TW6cmUlDfTYXCwuQkE/urOkgL1/OXpSNZ8WU5McFqWnqcXJobScFJIktquZT4UA33rxGdmtXhYd3xBr8TB5ElUt9pD2Bq2FweNp1oYlJiKHvK27h+fBw9DhePvDcQwthb0c4rN47hrX01bCpoYkF2BCNjgihtseDx+nhmfSE3TY5n5ZcV/mPmZITzwaGBlXM/GrrsjIkLoqipl6KmXv5+/Wj+tnnAydqcbh5bmM7Ggiaaux3MTg+jw+qkqKmXSUlKQrVKpqcOaKUsOyV8BWID5xsmxgOQYNIyO92MUaOgusOKTikj2aRFJT87/ZLvFEkzBuLO/Zh4F6iDBu/bWgJtZaJTM2cMvc/pwuOBg/+GbU+LfzefgPItcNtmiMgRt33TG4HWBFJFIJNFKgddhBgeUQWJf58pjElwyVMDvG+AcbeCaQjFxQsY9hjWjnbhwjMAACAASURBVHzRokVDbtcq5cxOCycySE1LjwOJAJsKmogJGZD3nJAQwr6K9gDxKoNGTn2nnStGR/PsxoHXykSThplpYfx7VyXJYTo2n2ji0QXpHK/tIjZExfS0MEqaerlzehKv7aogMlgV0BwBQKeU0TBEyKem3crSMdHMyQhjVEzQoNi1z0efbrcVp9vLyp0VLMqJoMPiZHe5uIo06bt4ZF4aEgGcHh/tFgcXZ4SzcmdFwFgjY4KobLUxOjaYS3Oj2F7UzENz01i1u4J2ixOjVsETH+UzMdFIWoSO9w/V0m5xMj9raH7yxEQjO0tbA7ZNSBwoL/d4fXxwqJ5VuytRy6U43B68Plj14/HMTDv7womvm/dzQtRYuOFt2PIHUTp28j2QvnjwfpW74c2l4OzTrhm1DC75rehQT8dGa7t4rC5czMJX7IC9rwTu43aIicZ+Rw7QUQn1R8Ftg7BM8TNBAGMKLHxOVEr0eUWnfe2bsPtFkdKXPBsm3R1QXHNadkqk4m+LGCly1fWR4nf+QBWM38mcfwcYrnYOa0fe0tJCdHT0kJ+FB6kID1Lh8fpYf7yeILUcU5/C3bQUE1eMieZfXwY6OrfHy6KRkazeWxWwvaLVypJR4qrmtZ2V/OayTN7eX0O8UYPZoOK21w9gdXpIDdPxyrKx/OXzE8xKDw9gk7RZnGRHDb4JFuZEYtQqaO5x8N991ajlg8M2GoWMa8fFUdbSS6xRw86SViYkhpISridYI0cqCOwpbyUnOpiXt5WRaNLy6Px0qtutbCxoQimTcOu0REqaeqhptzExyUhTt533DtURopGzeGQUBrWcMJ2CCQlGBEEg0aRDo5CxtbAZnUrGp8fqSQzVkhahR9YXWlqYE8lHR+op7Yu1p4brWHBSI4gOq5P1eaLkrM018MA8Xtt1To48YN67G6GlQIxpm9PPXvxJroTU+f+PvfMOj7LM+v9n+qSXSS8kpEEoCb0jTUCKgIKAumJdda1r2bWsrt2197XrYgXFgoAgHem9hBJSCSGk9z7998eZycxkAorivry/6z3XxcWVZ5555n7auc/9Pd/zPdBtJNjMUpDT2VrrpHDF5CZAdvBz6DPbK7no9WzarFC4CX56UJxynzkw5CbIXgZaX8+SeRCs3Gk1+fD5HKhzPK8qLSxYKkwVlVoqQGP6CZc8MAa+uRFqHBzs/Z9IJ6HrloO/9zU/2zuEzk8qFhP+II55YylU5QjzJjwdAs78TJx1nBeQXajjvKDTxLt37z7r5wWVzTy94hgWm52wAD2tJguPTe/FqJQwnlpxjFGpnlGURqXER6P0cDoAkYE64kJ8mNonCoXCTovRwoLhCWR2C+apFdkdUX1uZTNvb8zntnGpmMw2rh2egEalQK9RsmB4AieqWrh9XAqBPmqpkhwYx4hkA60mK29tLOCHg6WM7RnhITPhbBr8zMpsFu85xYurcxiYEMzAhGDWHqvgjfX5vLouD4OfDrMDJ5/ZL4bbv9xPk9HCw1PTuefiVLbkVqFRqUiO8Ke4poUxaeFkxAVR12rms50n2ZBdgd1u49YxSWhVSt7elM/uE7U8PasPTyw7wh1fHmDGv7exKaeqY2zJEf588eehfH7jEMe/oR79MgN0ajLjvDHV7r/QuPiXrOO+1xRKdPzZZfDFHPhokrBPfo/pA7p24gDGBqjK9t7eXOG1yevZLM+CL+eKcqHVJDoq294QLHvgdZ77+kdBjJscavFOlxMH+f6m54W2B5Icjc6AHpcIdbIm1/N41cdFUbEL+6V36A+zqhxYOB0+mwWfzoAvrxC46gz2PzbOc7QLdZwXdER+NqtrMXH/koNM7hPNkr0l1LSYuHJoN5INfqw8UsZ1IxKJCNDx0pwMCqqa8dWp6RcfhBIFp+vb+d5RaHPT6O6YLDY+2HKC9OhAXpyTycfbCimqaeOGkd7JpP3F9SwYDsOSDXy3v4Q/j05ibFo4/96Yz47CWu6dmMpj03tRUNXC5rwqvtlXwp9Hdyc1wp+8yma+3FXMQ1PSKaltxWKzMzLFwOPLPJ3TpztPEhfiQ1mDC6pZerCU1+f3Y1KvCFQKSfTuKKhhR0FNxz4z+sXQYrJistoJC9Dx8XWDKahsxg60tFv4fFcx7WYLOxyQTXFtK3ctPsBLV2Ry9+KDotuy9DAZcUEd7JLIQP0ZqYI6jYo7x6eyu6i2o/R/VEoYAxyStL/b8tcLBOG0plKJQC957o8RXfILh+QJULDec3to4tm/Z2oVKmFnadrsZTDqXijYABc/Lk7exyARvjszpLHM+5h1RWBuk2geZLXQXi8VlF3Zmbb/0WazCful9oTAT+E9BZ7J+spzcik7KGqOhttEXMvULBPab2AM/Z952wXtyAcOHHjGz4pqWogI1PPTkXKOl0vz4v0n62huM2O22nl9fR6Beg33T0pjkaPVmK9Wxe3jUpjUK5L4EFH+K6xsZoMjCj1e3sT2gmrBl3NrsOPNsU80+NJqsvDMymyqmoxM7q2gR1QA0zNjuHNCKkXVLdy3JIsQXw1XDU1gXI8ItGol141M5B/fH6GgqoVnV2bz4hUZVDcZqWk2dXSNd1qr0UpEgJ4ZmTGU1rex96QU8Bw53cCkXlHEh/jy14tTsdntbMqpIqukgUC9mr6xwew9WcuB4noeWXqEj64dRM+oAJqNFu5adIC5g+J5oBNX3Gy1U9ti6uhuVNFopNlo4dcCI33jgvjh9pEUVrWg1yhJiwzA0EUTh3OxjvtedsD7w+IdgjFrfpnGeM6m9YNJT8H3VRJha3xh4lOACo7+ICyRiHRQaTyfzarj3tAJyMQQPxS2vy5RtyEVpjwPsf2FyVJ2ABpK5Zjdx8AJN+LXwGs7imUo3gkr7oXKoyKo1WcOHPnGtW/v2WDvemI72zt0Xix/LXx1NVgdDKrhd4p++onN3vtWF0DeWjmXxlPQYxpc/BiEpf3x4zxP9pvHWXVcIDBTC8QPkZZ957HT0wXtyOPizlzBplIqSIsMYPVREWP6dn8JU/pEo9Moya9s5uoh3Yg3+PLK2tyOaLHVZOW1dbncPi6F4+VNXJoZw52LPJ1FRaMRP0cnmh35NcwZGMs3+yR612uUXDsikUC9hvgQX2b2i0EBHCiuZ112BWN7hBPiq0WnVnLXhFReWp3TUcgzZ0AcF6dHsC67kogAHc1GC9FBPujUCoZ1D2VXUW0Hn316RjQA1c3t+GpVPDo9ndfX5xHko8Vqt/PSmhz2nKxDqYDLB8SREu5HekwQ+0/V8dbGfP4+uScrD5eyPruSN9fn0Wa2cs3wBFIjBHOvb/XkfauVCsL8tTQbLYxINhB5jtrfCQY/Egx+VDcZ0XXOAZzeL5Fpay30uVwc2y/oVHfc9+QJcPALzw97X/7HOHGnRfaGBcug4ZQkFo98Bx9PlM8USrjiE+g1w/PZbK6U5g2xA+H0Ptf2iU9B8jjhfDdXQlCstG9rrYfVD8GRb137jnlAnGHZAXHWGfNle22hQDZOCuLu92DorTDj3xLl+oUJt7yuEGL7eUrGcvZ36HdbYxksu9PlxAF2vAm9Z0Lvy6RZhrulToBF8ySfACJ5azHB3IV/7DjPo/2mcVZmw8KpnYrRlkLSRedtXBc0Ru7UMujS7IJtB/qoMfhrmTsonthgPaF+Om6+KAmNShQK6zo5LbPVjs1mZ82xCpRKmRA6m3PVvvpYBWH+Ot65egCPTEvnyRm96RUVwPaCKvRaJeuOVZARG8SG4xX0jQ1iR0EtZquNWf1j+GJXsYfq4jf7S5jaN5rX5vXj5SsyqWxso67VxKLdUlb/7Kw+jEkLY8HwBAJ9NCzec4qeUYFszqvm1bV5PDKtFxWNbew7WcceR4Rus8M3+0oY2t3AR1sKMFvt9IoOwmK1cf3I7vz1q4MUVLdQ2tDO8z/lsK2wmlvHeLYKG5liwE+npqSujeFJoTw+ozd++nOb30vr23hjfR7T39zKNR/tZlt+tYh6lR6UB3jrqwKJfDoTCn+53KDjvieOkghPqZKb0vcKcRB/tPmGCCbdVgebX3Btt9uEPdJQ4vlsBsfDoUWQOFrKzqe9Ig4/YbhMBhHpQn80pMh5VGV7OnGQazTlebhjL0x4TJw+CGTRmUfu1Bs/9r2Mb8+HsOVliXg72Vnfod9rbXWu/IHG15V7aCiF9EtFvAvknPsvkIYTtk4Cb/lroLHsjx3nebTfNM4TP3dRjPYimH53zWSHXdAR+dmssLqFnPIm7p/UA7VSQUldGysPl7O/WJzcqJQwxvWMJECnpsnowi6VCkiPDmRAfBBqFFw9tBuf7nCxWHpFB1Ja78KmIwL0FFU3026yoVMreW/zCVDA/EHxKJWwIbuKCemRxIX4oFEp+HJ3Ma/O7cdXe7x53gVVLVQ0tnOstJ5Z/eN5YrkLG9+QU8k7fxqARqkkr7KJ42WNhPpJEU2z0eLoAmTpgFncLet0A0/P6kuz0UKQj5q9J2spqPIuwFl9tJx+ccH8bXIP2sxWDH5aDH4a/HQafrxzNFq1gpoWE4VVzSQa/H5Vd3mbzc6Xu4s7tGbKG9tZ8PFult0xkt6FmwTndbfNL4kGh9bX+2CdLSBSnNqABVIwE9z9v4uptlR7b2uthaZydDo3+CisB8z9TBT9vvuzK0Id/6g0JdD6SeRZdVwcs1IpmiYnt7uOYTXJRNFZR0XfRYGOSgvtdTK+6EyBWkoPwL6PIWMuxA74/ef+aywgCqIHQNpEgbtMzRAUL31Fg7vBrLcFZlEoZCXifr5O84+Q6/P/s7XUdLHttxejdWUXtCOPiYnx2lbV1E55QzsxwXo2ZJtpbpeXps1s7XDiII0bRiQbeHVeZkcn+qUHTjNnYByvrsvlH1PTHcnGMHpEBnCguI64UF8y44LJKW/iqiHdiA/1oc1sJSpQj0al5O6vDnbAHxuOV/Lw1HS+PyhKf5GBOv51eV/2nKjjq92nGJQQ4uV0fbUqYoL15Feq2Z7v7SR+zCojr6KZqmYj91ycysrDriRYmL+W5PBo2s02TtV6OkeDv5b7lhzi4anpDEoMpabZ2GVRjsFPx7Gypg59mbgQH16YncF/tp1gwYhE7lx0gPpWMz4aFc/N7svUvtFdVrm6W2WTkU+3F3lss9rs5Fc00buzHCtI4wN3bZEuzOO+qzUQ7t384r9ioUmyGnCPIiN7w/5PuTxjnKv9mkotHOwfbvOEGTY4oJXYgXBsKXx/i+vcB90gydFTDhZEzAAITvAeQ3ia8MR3vu3aNv5ROLpUHHqf2bD2Uddn+/4DN/wk5f90/Q6dN/MNhSnPwdd/EujIaWGpcj20fhDlJgsb1UcooLk/yd8KBUx9CQKj/9hxnkf7TeNMGuu5sgMYcuvvKzTrZBe0I58+3bNg4+CpOu5cdIBTtW0E6NS8eVV/sksbiQjQUlLX5tWibU9RLZtzq9h5opZAHzUvzs6grLGduYPiqWs18+zKbCqbjIxINnBZvxiign14+LvDVDUb8depMZpt/Pvq/oT66fh42wkvFc7tBdVkxAZxqKRB+no2m3hyVm/qW0zEG/z4x/eHKahqEcx8fCqbcysZnhyGxWZDpeoC0kF0XWpbTDz6w1Hun9SDvMpmMmIDAQU3f7aPh6b0ZH9xHRWNkiAd0j2UmmYTvlo1VU1GXloj1LT7J/UgyEfToYOiVSm5KC2Mp1a46HVT+0bz9I9HeXJmX275bF8Hdt5mtnLf14foERVAz6izF4ho1QpC/LQezZIBVEqlCCP9/JynIxz1V9D5czbrfN9/0cxtspzX6F2QxPmw8J4SaS+/W3DoyD7Q7ypY9xgBh5dAfF9X+7TWWldnGXdrqhCseMU9nhPY3o/hkudFkzxhJIx9SCCdzqYLgIv+Lg6wqUwSpkqVjO3EJjjWaalvaRf1QYcjP+drea5Wm+/pxAHWPyXn1Lnc3z8CLn1DmEhtNQI1Rfb974zzPNlvGmfsQLhysVsx2u2Qfn7P94J25B9+KPrLADXNRv761cGOaDTET0tueRNKpYJlWeUoFfDw1HSW7D3VId+aGunfAXH0iwtmzbEKvt0viUtfrYr7J/XgxdU5bC+oITbYhxmZem6+KAlfrRqLzU5ssJ6v95aQU9FEaoS381EpFVjdvLvVbqeqyUiATs26YxX0iw9mqqNH6JqjZYxMDWdvUR2xQT4MSQplw/HKjslBpVSQGR/M8ixXFB6gV3PFoHjaTVZaTVaSw/15ZW0ufxqWgJ9ORWywD7tP1PLFrmJuGt2dz3ae5MZR3fHVqrDb7dw3MRWLDSqb2hnQLYS8ikb8tCrazFamZUQzOjWMqX2jwI5Xw2mLzU5ZffsvOvJQPx0PT0nnls9dSb7JPUIZGW2H4Ay49kfBdFtrYOgtws44h/veYe1NUJMnEX1osqv4paYA1j0Bx5cJDHHJ89Brxjk1/j2jqdTQcxr4RUDeTwKLrP2nRNJWsyREnY48IErghHo3LQ6lSpgnpQcFduhsNrPARqZ2CDmLbopviGDs9cWw9jFpqaZQQN95kHKxZ4IVBMZxWJfX8nyae/GU09rrvZtjOC0gUv51sj98nOfJftM4NXoRQ0sYKc+Nn+GXv3OOdkE7cpvNFcFUNBopqnaVxU/qFYlSqeDpH10R5sacKv4xLZ1nfsxmSGIoNjsdEeng7qG8vMZVSNFqsrJwexHTM6JZsq+EHYU1XDEwjjA/HV/tK6ai0UhqRADLHF185gyMY+Xhsg7tcoUChicZeGal/H6ATo3NZue7/afJKW/i7otTvUrox/aMpK7VxJVDkmlsM/PKFZnsKKhBqVTQKyaQD7YUeuzf1G7mnU0FHb/3j6np/GvVcT5yHPfGUd1Z62jYnBDiy61jknlrQz41LSaUCrhmeCKDE0LQqpWUN7YTE+TD8jtGYbbZSAzzQ6cW+OVUbatH9A6SS4gM/HU0wrE9wvn6luEcPl3PlMgGIo98gOqzDbKkHHEXzPmPo7z81z1u7vcdEHbEun9C1tfyd0RvmPOxOPStr4kgFUhE/P3NIo96Pjvi6ANh+1ueAlMKpdALnRYQCbM/gm+ulyIgXaBEn8YWaW3W2cmrtHI8/yhJZm54CtJnCDXtTPmD4yvFiYMsPbMWC6de6++aKJQqjyrUyLAQmezU53m14rSYAd7w07DbuqwyPZt53fML1H7XOP9A+YML2pG76/4G+qh5aEpPmowWdGolscF6lh30LqQ4crqBZ2f1QadR8ZAbZ9pdPdBpxbWtzOwfw5yBcVyUFk5JfRubcqrQa9TcPSGB+5ccJMRXQ2O7hc92nOShqenkVjRhtdqZ1DuSE1Ut9IkJIjbEhwHdQnj+p+NcMzyRrJIGmtvNTOsbxY+Hpd/nFYPiSAzzY87AOPafrCM2RI+vVsXwZCmZN1pslLsVAF2UGkZBlSvasdthXXYFQxJD2VFYg06tZFBiCL1jAlEpwA78e2NBR2Rts8Mn24voFx/MI0tFW31mRiSXDfQucY8P9eXlKzK5/cv9GC02VEoFT8zoRXJ4p1VITYFEl+Y2wTujMkCpRKdRMaR7KEPCLbBwvkse9dAiwYCvX9VlFHYm89J7Lt7hcuIgfOo9H8LIuz351E6ryvV05KYWgTAaSsAnVBoch6Wcuaiotkgi8IINkDxRWCcj7xamgd0GCgXtY/6JPswNu68pEMhg1H2is2JIku9lr5DimAmPwe73hU7oHyErB1+DcLCNji6Jez+CeZ8L46Oz2e2Cs3e2wp/hykUyNn0IDLtVlvIAtYVMbl8Gb90kMMfkf0GvmWdfrdQUCgXS1CpQUnSGF6XRw2L6i677z88L9DPklt/ELPrfoEUOF+44L2hH7lzC2O1SxfivVa5O4A9P6enR59FpNpudj7YVYbHZeGhqT575MRuLzY5PF/v2jgkkIdSX1YfLO7oNzciMIcRXgwI7z12eQV2rifpWEyV1bbyyJpc3r+zP6mPl5FQ089HWQgbEh3CypoWfjpQT7KshyEfNXRNS2ZxXRYBOzaPT02lut7D6aAVLHFrfMUF6Hpnei8KqZgYmhFBU08qm4xX8c3pvalqMJIT6Utdm4snlnuXi7SYro1LDiA/xYViygYqGNlRKJVabndRIqRztbEXVMhloVUoWjOy6K09pfRt+ehWf3TiEFqOV6GA9SWH+aNVuD211ntAHGx2t51QauOYHSBzp2qe20FvjurZAcNQzOfLqPCkqqTwOqZMhYbj30rW0i8KggvUw+j7BWTt3jXcXuDI2wbbXxdGBOLHxj0JzmbBnOltrLSy7Q/TLAXJXSzeb8FTBsa1GUOnQdx/m0hVvLIXFV3uW94+8W37HkCqwybrHJTHZ53JQqCFhlKwknE7caZueF/ipc/SmUMh4i3d4bo8bJNu7jZBVgtPRWC2w4220x5e6zuv7W2RlcKYuQDX58NnlUO9gcSnVopfe1XVymlIlsE/cIIF0usL5nb+vVJ8xKv3fAKvAhTvOC3N6cdiKFSsAOFnTyj9/OOrx2bOrjjNvcJxHUKVSKhidFk63UF/unpBKVkkdd4xP4Y7xKQTo1Dw2vRc6h3OKD/XhoUt6sC2/mlw3B7jsUCkZccH87Zss7lh0gC92FdMjKpC4EF9uuSgJi81GRUM7dc1Guhv82JRbRXaZvIw3jeqOySIFO7tP1LL+eCVvrs/HR6v2kMUtbWjn55xKMuMCwW5Hq1Ky+lglj/5whNN1bby8NheL1TtgnNQ7ip5RAfSIDuBwSQPtZjuPLz/G48uP8fXeki5x/F4xgTxwSQ+W3Dqc2tx9Xp9nlzUy+53tXPn+Lua+t5NPtxcRoNN4OnGQ7i5OJ+4XLkvqne94cmHVZyjUUZ8hAqwrhs9nS7ux/Z/AV1fB/k/ZsL5Tibw788FpyeMlsp38jMAUTksYJVVzTqs45nLiIKuJ7W9KQrDZpSuDqVV42BVHXE6849w3yzlvfEbokxufofG0m95JZba3RsuOf0uZfUQPuPR1yJgHp3bKb/acKpNNQBRcdL8UA4U6+P2WVm+utdM6l/aH9xTaX8VRga3co8WWSjj8tfcxzqZVU7zT5cRBcgEbnwVjF/h+Z9P6de3EW6phz8fwwXjpSHR8pbSE62TOd/1Ctwt1nBd0RO5sXNrQZvYSugII0Kp568r+bMqpwmS10Sc2iDVHy6ltMfLgd1l8sGAQeRXN1LSYCPLV4KdR8vr8fuRWNBEf6kt1i5kdBd4NB46ebujgUB8vb+LdnwsI8dOyPrsSjUrBK3P7sbuwhmFJBsb3jMBmtxMf6suJymY25nrSCg3+Oo6ebvD6jSOljRj8dBitVlLcIIxQPy1VTUa+2nOKf0xNZ+2xCtrMVuYMjGP10XKSwxPQqVUolQre21wghTfAT0fKeXBKTy+MPNRXw+Sx8vK/v8oTgzdbbHywpdBD02VjbhV7T9YSG9IJT20qk5ll1L2Cx1ZmgyFZEn5OemBYKgy4DvYvdH2v31WyvSurOOLpOAB+fp7IMR85BmiEnJVwag/0vBSOL5ftEb0cXeJVUoRz8yYRadIFCO0twE2WtwvBK5rKxOkZG6UHZnUerHkUclfB+H92PVZ3UyipU4TSEVt21lgBsFvlHwjuHTtQom9dgETO2cvhu5vEEas0sro4vARG3isO0dgsCUN3gS+Nr0xgfWbLvWgsgy0vQlia0CLdTesvErilnSbvs2HXXfHmG0oEy/8FptEZLWcV/HiP6+/FV8K1K6D7aI/dnO/6hW4X6jgvaEfutJhgPfGhPh78aZ1aiUKp4M5FB5jSJ5roID3PuCU+RyQbsFhtNLabUSpg0a5iYkN8qGo2Mn9wN/7+TRbJ4X5kxAVRfswzQggP0NHolvjbX1zPHeNTWJ9didkqCc3bxyYx572dKBSgUSoJ8dNw+9gUtGrPMPpUbStXDY3nh0OeD8CgxBA0aiVv/1zAS3MySI8OILusic15VczIjGHJvhL+teo4Q7qHYvDT0mayUlTdytaCauKDffHTqjyqVo0WGy+vyeXZy/uQW9GMWqlgzdEKZvY7M++1xWRhT5H3RJZb3gT1JeJgnJBI4ihx2rmrxQGD6GlU50oPSX2ARGXjH4a0SVB+FKJ6QezgrjvfQNcO0GZ2UTPLs+Cb64R6FztAInO7TRxOm2PcCoU4sc6OzGnB8XjxUg0poNSKUzO3w4anxYmDKAnG9PeEcxJHg9nBwvCPgGmvsvNgIwnDHJ+Hp0vSsrnc9Z2M+dINCKSZb9khWQ1E9JSIe+lfxImDMBm2vAzzF0mBTf4G2PSsUB6H3iqyBAGRAk0c+VYYQO6m7iIpbWqRaH/JtS4GSdxQWUmdyeIGe28bdKOXFnuHGZsEwukqErdZZYJsb/AWIstf5+XI/89+n13QjnzmzJkAhAfo+fdVA/j7N1kcL28iOkjPC7MzOhoZ/Hi4jEszornn4lRK6tro3y2IMH89z63KIa+ymSAfDX+b1IPPdkrjiDfW59FstHCopIEHp0g3oJI6mSSm9IniZG1rBzsFBIZx7xNa3WzkUElDR9GPyWqjotHIMyuzeWFORkcnIZDAL8xfx8x+MSw/VIrNLolMH7ULsy+qbWVYkoHJvaNoN1sZnRqGxWbnpyPllDe0c9XQbqw5UsYTM3ujVysI1GvILm8iJcLfqyNRUXVrR5XltIxoj2jfeT2dFqjXMKlXVAcLxmmZQS3w5ihJxl3yPKRNgdhBgnMe/NLzJuWsFBw8xgFn+EcKZa/nrxDgj0gXyqC74NSgmwhPdjibGkd3o7jBwupwt5M74PofJWJtcLRDC4j2diphPaSh8cq/g7lVKidH3C2Yri4A6k66WC8gSdVR90DqRCjaJpNI+qVyLfrNFzZKYDRjQytkcmitETrZNd/Cvk+hZLdEzL1mCu2splCi0CpHfkcXCNNe9qYjWs3iqOsK4IvZLs75Tw/KhJcwSn7n4idh2e2u7wUnSELS3VqqYNldUHFYnLnNIjh90jjp1Za0RAAAIABJREFUneluphaZFH1CZbKc+5lQLNvqBOfvd5X3fTO3S9n5xmfB1CQJTr8w2R4/BHzDJHG7+QWZrJLGyYpjy8vyfV9v+l3nZ/NCtQt1nBe0Iy8pKSEyUiLCjLhgFt08jKomI8E+GiIC9RTXtHDPxamYbXbaHHTCiEA9w5JCeW1dbkfyr6HNzKPLjvDgJT1paDOz6ogrcnplTS7zBseTGR+EzS6J1W2OqkuFAhYMS6B3TBClDa28OCeD7QXVdA/z47V1ubw0tx+mjflklTQQ6qfl3olp2Gw2HpzSk9zyJlQqBcnh/jy5/Bjxob7cN6kHUYE6fjhYisVmJ9pPpEfD/LXUt5iwWm0MTgwhp7yJcH8tz17WB61KSX51Mz2iAnl1bS6XD4jl+VU5TM+I5v5Jaby8Rs4zxFfDEzP7oABuuSiJHlEBDO0eSqCPpsvrCaBUikTB4ZIGdhfVolTADYMjGFC8UBxLU7lEdDesgW5DPel27vZbJWXDUuHaZbD7Q4m+M6+E9Espyi/BEBEp5d4gXXM6W9UxcVjF2x0FO45y9Rn/hmhH553SA4Ljt1TDFf+RZg6+YRKlO4tVtH4Q3V+O1XBKtm19FeZ/IUnJhhKBYoLiIdzVBs1eXwxHP5LGE4ZkGPuwUAGtJk9Rr4INLicOAue010uLNncNFbVOYJ789d6Vr7vfFxw8ZxXM/VRYIie3QlA3WSmFJHruX3FUrou5RZwtyKR161bP/coPCwe/aItwnC9+TDj4CSPlWpzcJvmFtEkQP8yVqDy9V4S8nLbq71KKv+tdgXQue9dz4i3cKBNIaJJMECkTvG5n52fzQrULdZwXtCPft2+fh2xkiK+WEEcTX6PZyraCGv69sQCT1Uawr4b7JqaBArRqJcfKPNkAdjv4aFVsyavu6KEJYLLa2FNUy7CkUJoc7JJJvSIYkxaOSqXkg82FbC+oYe7geL7eewqjWbD4jPhgcioamTcojrsnpLIlr5pX1uZy46juvLo2l4emprP0wOkOpkpNi4lJvSNZkVVO37hgYoL0tJktvDYvk7yKFlYeKSdIrybe4EdMkJ5dJ+pIiwrgdjd1xnsuTuWZH7Ox2eG7A6dZm13BzaO7Myo1HI1KwR1fHuBkbSuBeg0tJgsfLhhErFv7u87XEyAp3J8Pru5DcUkpWmsLiXueQHdyk+eNqDoujjwsTVqmueOuPaZKJPxbLToTLn3N4QB9wGYl1r4PflwsEfSchRJRjn0QDi5yYepxg8URfX+zKzlYdgh++AssWC6O+ZMZ4jgj+0iTh9jBMOFRTz57Y6kU1TSWSPIw6yvwi4TjP7pWH3s/hOtWQjcHlmI1E3z0U8j61HGM03BqF/x5ozfE41xVuFvW19Ivc+0/xZlr/URo6+cXu+5grw92TQCLrxQ1xU4dizqstVYmrgHXgI+DQ571lcAg7knGpnJYdBU0OLjtBeslYXv1t/I7R76TXIfVDHs+EE78wGtl3/x13r+bs0ogqJyVch/UOhd0BALDTXlRJp3IXl5f7+rZvBDtQh3nBe3Iz2a5lU08/P3hDuizvtXM25sKmD0gFr1aRbi/zkvn2+CnZd6gONQqpaPpcBP3TUzjRHULTyw/Rt/YIOYPiefuxQeY1jeaPrFBHC1t5OGp6Ty7Mrvjt7JON3DfpDRig31oN1u57Yv9HTz1L3cV88rcTBbvPsWlmdGMSQujqslEfKgvBZXNGC1WFm4TKOPtqwewraCG9zdLErK2xcTfv8nizSulgYR7EhLAbLN7QD5N7RZeXpvHRWkR7DtZT1GNFEw5C3teW5fH4MRQ/HSdbrPNBqX7IecnsFsJ6jaUvqv/Lgm5ti6Sgz4OuMI/HGZ/IPKjBRvFCVva4LtbYOrzImrVWiXR35lw8a5MqQKlg9lyeh9RK6934ecHP4dJT8OWV2D8I9KrUh8EvWcJK2TE3YJfqXQS/RbvkAYU1XniTMY+IzrQrTXS2qy20JWcLT8M/5niCXNcsVAqONc/4Xa9rHD0e5cjbzyN/kgniMlilARwZ0ceP1QiVXeLzoRd70vnoG4j5Lp+c5NI0Y55QFY+LQ5GjUIhQljrn5S/TS0y+Ri6oJJajHJ9tr7q2pY6CdImC5btLshVW+hy4k5rLBWYadNzsvq4+AnRcbFZ5XqkTZZEchfQCD7BrutobpP74e7Io/tLJN4Vnv9/9rvtgnbkQ4YMOeNnZfXtXtonZQ3t9I0N5uNtJ7hpdHdeXJ2DxeH5rhgYR3SQnrKGNkwWGzePTkKlVPCf7UUcPCUY7frjlWSVNPCnYQl8tuMkEYF6YoN9yK1o8vqtDccruXt8CsuzqlErFVx3URJJYX7469X4apWM6xmBn1bF13tLuHpoN55ccczjGD4aFXY7HT0v3W1/cT2DE0IYkWwgzF9LdbMkqzQqJUoFHs48xFdDi9GCv04w9/HJ/szubsaGgo2VaixulWgd1/P0Plg4xSXwpFRJhLj+SSlccZahgyTHHLodgDiQbiNEY3zfQoneFAphlmx7TQpgIvtIR5z4M9+/M9qxH7yToPnrBb/d+qqILBWsF771jDdED9vJxc68UjB6m1UglCkvOJKKjgnx5DZHFxuHIy/c5I1V73xHIsvO5n7zVFqZTDonHbsqtOk2TGCHnW/LOPpcIcVKBRtEFS9hhNAb6xyMom2vCzZtNQttU6ODA5+5EpZaP8Gzi7ZJMjoszQUT1RTI990tb41g8uG9PJknugDvJDAgij8IC+jwEllxZS+XsTvvS9oUOPytq/GHUiX7rXlEjhk/TK6tk/PuEwIX3XdWJ362d/03W3Weo8ZAIWymM7GnzsF+cZxWM1Qec3RMCpeJvbPmzB9gF7QjDw8/AyYLRAV5c5YjAnT461T0jg3is50nuXN8CmabnTB/HRH+Wl5YnUOYn45pGdFEBuo4VdvW4cSdVtVsJDrIB4vNToBOjcliQ9+ZUw3o1SpWHa1gSPdQZmbG8Ldvs6hoNKLXKLl7QhoBOhWNRgsRATpqW0wE6j1L4K8dkcCqI6WE++u81Ax9NCo25VZx5HQDb145gK/3FnOguB4/jZInZ/bhieVHMVvt+GlV3DYuhdu+3M+1wxN5bqKB6eXv4r/1e1AomJy5AG1bJBhV4BtKdLQDAsn6ylOlz2YVidGwHhI9jn1InFJod4jK9CztNjbJErzskItSFxQnkbNTpvTkNumzefOmc395LF1odNgsoFAJvlqVLZBHxjxhm7gX1BxaBJe9Byv+KnCJ1eRZVg9S0p8+UxKHzp6Y7mZskih2y0uubQqlFPI4LTAG6/jHUa2407XNkCKNkU0tcgwfgyg32u0C1Vz2nlTFFm6QatRpr4jz0weC1m31YmmXKklfg0A1J7e7el0q1aJzvuRaF2STNgWmvST3wNLWtbKkqUUSj7GLXfi9IUW03re/4dovY57g2U4rPSBRdPZykVpQqGDfJ9LsI24QjH1AnGVwN+HNR/QW6CphhGD5FcdkTOE9XDz5M9jZ3vVfZc576ZQ3KMuSXqFOITPfUIHcuqpJOAf7xXHmrIQl17mpXN4owdF5VDrsys5LQZBCobhYoVC8rVAoHlcoFI+dj2MC/Pjjj2f8LDXSn39O79XRGMJfp+aVuf0YmRrO3IFx6NUqXl2Xx+GSetRKBbd9eQClQkGP6ADe21zIK2tzsWPvsogG7Lw5vz+tJgsLRiSQGObnURmqUMCE9AiW7D3FA98eprC6pUONsN1s4/mfjqPTqFiy9xQjU8L4dHsRj0xLZ96geIZ0D+Xeian0jQ1ifXYVl/WPRe2m+x0ZqEOlVLC3qI7EMD8W7T7JC7Mz+eH2kVw7IpGEUF9en9ePO8ancO2IRN7/uZD6VjPv/lzAkCgl/nnfO07BjvbgJ3BkCXy1AA5/Q/iOp2DzK123JbMYJXqsLxZHEj9UHI3TidcXixP8aBJo/GDMg7J8NqRAxpXejXUDokT5z9oFxdBpTeWSTHTXr+hzmXfyNHWiYNAhiaB2vKghCVKJ2NnqimTFoVDQEV26m7NJBYjErKLTKzDybmGBXLNUmCcZc+G6FcLacbMGTbi8oEP+LMyQPnMcuPOV8PYwWH6HSAVU5Ur3+CXXyuRwer9MTD89IEnG4yslaZky0XMcYx8SB9l3jmDiV38Lf9kBuWs8cffcVRKdgzBTojoxWPwjhUvfWuvitTuv96h7JXF66RvieI1NUhTktKB4QCmTzsDrYNc7sPwuuRd7PoAfbhd2Uu9ZMP9LuO5HgV/UDmpn8lgRi/oFJw5nf9fPasZmWQV+cqnkRI7/KE49a7GnGmVrrXczj99gZx1n/SkJIjxULj8SSYk/2H53RK5QKHyBd4HedrvdqFAovlUoFBPsdvv6X/ru7zEfjZo/DevGiGQDNS0mYkN8SDSIQH3P6EDeu2YAJXXSUWfR7mL0GiXDkw0ewlnbCmp4ZW4m93x1qGPbiGQDe4rq2JKby32TetArOoCGNgsvzsngcGkDze0WuoX68uWu4g6II+t0A3EhPh0URoBTdW0MTzKws7CGtKgAjBYbu4tqiQjQ8e7PhSiA+yf3wGi28dZV/ckqaUCjUmK22njv50Km9I3iaGkjsUE+qForCGoqo1oRyp2LclkwIqGDYug0s9WGvasmvqf2wICrYcXdKJ362ROfkmWzu3W/SBzMxU9J2b2zOUFrnSQdNz0vUbcuQMrb3VkJuT8Ja2PdY3L8MQ8IrW/FXVLe3v8aiQp9ggUKsJrkhTu1SxKPgd0g4wphNsQNpnraQsIKvpXfTRorQlGBsSJKFZYmLAqVTvjWp/d6nofS+UgrZFWhC/CM2sc+5CqyiRkAC36Ara+LrOqw28ShKjXi4DP/BCqVnIvCoTnuiPjUdQWw6QlJRJpbYP5iofvFD5brl/WVlPpf9HdZ3neu1jS3Cbtky0sw/C5xpmWHhCETnibVqQqFwCeRvWR1VHpIpGs72+l9kDlPzuvyDyTqN7dJBG1IFsXEMQ8ILFOWJRK6pftlkkqfKYnT1jrPPptqnYyp2zA559oTAju5W2uNwAiGZMH5/yfs5HbRq3Ha4qvgpnVdV7BWZntvO59mbPTsBOS0rgqtzrOdD2hlOHDSbrc7MxvbgGmAhyNXKBQ3AzcDREVF8f777wOCOYWHh3fMdN26dWPSpEl8+OGHNDY2snDhQq677jqWLl1KZaXoHs+ePZv8/HwOHRIHPGLECBTNAbz/7Rfo9T6k9BuGLiCUB749THp0IGqlgmFJBtYd80zk2e2wt6iO9/40gD1FdQTo1ZQ2tPPVHqGh1beZeHF1DvOHxPPI0qOMSg1Dr1byxS7PJFGgXkN7p8rTEF8Nx0obiAjQM39wHHctPkhZQzsnql1CWGG+Kib3juLQiQpOVDbzU3YFdjskh/uREuHPDwdL+fvYGJTvjgKribC4wSya8hdWNyq8em9elhlNuMWtR6LGR0rmu4+WIh4nFmq3ixO/9A049KU4mJ7TZPlcncuxMR+ydeU+Lp8Vg1/5LvSbn0HRZ47sC+Lo9n3i+QRYjNgsRpT6YBhxpzjp0v3yWU2B4MGpE2HvR1hGP0h79BD8m8qEi2xqhj5zaM/fzLJ8BXPnzmXRjlPExV6GT6CeccmpNAX04kSNkazVhxg1KgC12pcNGzYwfeD9RLU8iKK+SBzP0FslQo0dKNz23DUw+n6oLcBmbKEufiJ7ipSk+hVgsVj4+eefUSgUDB/yCN0Tu7F/13aSWtYT7gNaHz/IWyddd+w2CO+JfcwD1DW2kN0cQHqwg4LWXi9RaHkWbHhSrq9SJY5z/2fYa/JlXaDx8eyW5BMiui0AO9+iIHAo9dpo4roNYMWKFQQFlRMQEMDkyZP56KOPmNU3EMPOZwVfr3aTBwDawjP54sMPsdlsTB87iKic1SidDl+pxn7FQrYUQ3jLKtK33u7C9ssOYTx9lPZxT3I8v5BG+0h6TR5JiN6OPSSR77cco3XD5yQlJTGuXxIqpQo6zUfVNbV89/77zJs3jyNHjnD0qESfY8aMQa1Ws94ht5CWlsbgwYP54gvpwRoYGMj8+fNZvHgxjY2NNDY20tLSwp49e8jNlfObMGFCx30C6N27N3369GHp0qUYjUaSk5OZULmQzmbd8S6tyZcS4A4TAXXdJpK3ezfJycl8+61E5xEREcyaNYuFCxdiMgmsd9NNN7FmzRqKi+U9nzZtGlVVVezevZvGxkb27dtHXFxcR9u3mJgYpk+fzrbDJxgemYGywk37R6miqFGJT0UFJSUl7NsnjK+z+T0ArVbbpd87kynsXsmOczOFQnElMM9ut89y/H0TMNZut//pTN8ZNGiQfe/evWf6uMNsNts5q41tyqnk9i/2c+Po7ryxPh+tSsm9k9L4ObcKm83OrhOeM+atY5LQKhUYrXZ8dSre3VTYIQdwz8Q0Xl2by/vXDKDZaOXJFUe5c3wqz6483lEa76dV8fbVA7j18/1YbXbiQ30Y1yOC3jGBPLL0CA9PTSfMX8ddiw94KTB+d9sIBnQTRkhjm5m8yiaqm4zkVjazKaeSP4+IY9T+e/GPSBR+bsURiEgnO2IaW5siOVhcS25FC7NSNcz0PUJcQjL89A/ofalg4OZ2gUgKN4qWibtlXCkOp6lUdFSsJon+bt8j0d+p3fDxJHFMo+6BvQth0PUSEW9/3TO6UaqEJqjSyXF+fs5V/em0MQ9KBG5ulch3ybWen1/0N4lMNXpsSvWvv+9NFQI12KwCtbTWSJJpxV/lc7VeKiatRtE5ie4ruL9SKdenpkA+0wdLsi7HsXT2CxPoYfXDrt/qMUWaKMQNxTbiLpSbn5eOPFNfhFUPeC6pNT4w7HaBmLa8LOe85WVx/L4GOd/NL7giuD99J9t9DcIBz1kDcQNkkghNhkXzZeUz5gGZmMscRWf9rhZxLmehT85P0uDY3QwpMPtjudeL5nt+FpwgXP7OXPTOZrPJ6mHjM65tgbFw/cpf/u6vsF/1rtcVi15NXZGsGOOGSsHUwc899xt4g1QYH/hCePAKBVz0gBQ3/U4t8F8cZ/lhKcYq3S+w1qWvS/DzKyWcf8kUCsU+u90+qPP283H0SsCdaxbo2Pa77cMPP+Tmm2/+1fuX1LXy168O0mKySocahCe+IquUGRkxBPlq2e3WrV6vURLso+W5n6RgI9RPy80XJfH6+jyGJxvIq2giPEDHlrwaQnw1vDG/P3Y7fHDNQPYW16Gwg69OzatrcvjsxsHkVjSzv7iOyCA9apWS28aloFUrMVkt3DY2mVfXubDNq4d2I80Nnw/00TAwIRTaGxgTWsNNCeBjzAKtQpa1zhLn/HWkxe1meehjPJhSQ4hqGf65WwSf3aaWhM4317k0Rna9A7PeEQElZ0SoUMhyesVfZQl/8WPQVi9Kff6R4txaa6XsvKlMovVpL8pLY24TB+cOrUx4TF6Y8iyBNgZcK44lb61rH0OSLPk1fpLQC00SCpzTspcDSmippCF2PCF9JwrHur5YqiENyTLRdDadv+y35RXBgNNniJLi+EfEUSeOEq65s4uNSgMLlolz2/wS7HkfUMDkZ11OHGQ5nL9OimNOOjDoU7ukofDudzkRPJLkSU8JbFR/0jvJaG5zYMN2uYbbX4f+fxIHH91PYBcnhhs7QJgpdUWCcTsn3aPfwP7P4JrvXMJgp3ZLmzhdgEycyWM9e152pS1TVyTn4i7t6x8Jw++Q5PGWlwXjjx/qWczkbkql/K4hRe5VZG/oOf28OHH4Fe96U7lovbtDaZOekcYch92S9yqNw2GHSzeqvnMAxXnTYv/FcUb1lfxKU6kwmwL/Oy3szocj3wEkKBQKnQNeGQm8/Qvf+UOsutnUATfsOVHL9IxoVmSVceR0I0dON/LwlJ68OCeDA8X16DUqEkJ9eW29y7nWtpjw06n426QenKhp4fsDp/nb5B68s6kAraN3pVKp4LV1eYT4arBY7TQZLVyUEsYPB0v5bKcTcjlN35hA/joxjcOn6+kVHcTEXuHEh/pS2WjEbLNRWtdGdbMJf72bczI2wabn0e/8t/w97mGBCTY953GeqpJd3DLKSuC2t1Cc2uV5EU7v8X6Zt78JV3+DfeurKLCL8NLehTD3c4n+1jwqjkgXIM5w1YOSJxz8Z/m8qVzoZE6sr/yIOL5Tu+TlL9nnkpK1WSTBM/5RlyNPmQjV+cKCsJoFUnFylJ3WY6rwvMtOE6LcDCEhsPRWSYaqNAKR6AJEMCtuiOxrMcpYvrnBxU4pPSDH3vCkiExp/T1bkTl1TUbcBbvfk2Nf/IQ4285WfliicKcjj+4ntDy7DRVWGU/cQJm8OkMnvqES7YaliszB7vcEyx59LwTEiv7IqZ1CdQxJlFXM6Ps8OeAgjrbiqJT9B8XJ6mP5XYLhD7pBft/pyM1tXVfZ9pgGuY6CprA0gWaG3w4bn3Zxvfd/KgnV1Iu9v+80vzBh77gzeM5kVblyrbS+4vTdhcx+i1Uc9c6HbHwGbtsJ16+WiUqhlADFXU+mczPr/4b5BP3hLJXO9rtZK3a7vRX4C/CGQqF4Gsg6X4lOrda7E/zZLNxfR4ivOMat+dX4adW8fEUG909K48EpPdh7so77l2SRHh2IAnhvcyG1nVqcma024g2+RAXpuXtCKl/uKqahzUyAXk1KRAB2OwT5aKhrNdPNIHK5N4zq7oWbHy5tpLHNzPCkMIYnGXhpTR73fn2I5346zstrclm05xS7TnTiIVdmg9OJg0Q+hq4bD9e3mlBEZnpuVKo9aYVOa6sDXQBb424Tdb/6EhhyoyTnnM0SQCaSjc/Ii9reIEvplInyMlTnCY1s0tOSAKw4Krzt3pd5CiI5zW4TJcRL35AqzMKNMO4RecnsNqg7IVEhCHSgUApbptdM4Z83V0iThrhBck6b/iXO+rNZkpTLXw9fzIXVDwpM4a6ZXbRZuO8hiV0zdBpOuYphBl4nlMuuCpiSxrnEswKihYp3chskjKKk1W0CNiRLdNhROBUpyV+rScSuDCky4Wl94etrhcUw4y1ZGVQclXOz2+UadEUfNDbJykrr65oc7TZpruFkrICwJna9J5Oos2gneTz0nS2rlIAogZlmvCWQlHvBDggHvfO232Ile+GDcdKU+fPLhTVVX3zWr/ziu+4+STrN0ibjjRsoVMgxf5PA57dKRvwKO1ef9N+y8wLc2O32tcDaX9zxHO266647p/1jQ3x4bX5/bv9iP81GC0sPnmZy70hO17WSkWAgu6yJ5HA/iqqbGdzdQKCPhlfWuhJHKqUClUKJ1Wbjqz2nqGoyolIq8NepmTsonhM1zXy8tZBnL8vAZLNR32qi1WjlWHmjR5GO04J9tQxNMlDfavLo9uM0d5YL4J3xLj8MOj+MiePRFW3o2GyKyGBXo4GEwdfBwU9cL59aJ46vc+ut3pfB1tcZPfUFqRgc9hfBWI2NeFlDiSeOaGmD/tc66GsmwZHVeqHplewRSCamvyf/GBwd3i+Hn18QXRCQ/Sc+KYUkAdGiTdLeIDS3b28Up7/mEYmOVRpJXmbMFyfQXOniCjeVwsr7XAnc0/vFeZ3aJddCHwy9LpNr0VIlhUruNvgmKaoB+b+5UpzukFtkNWGzSNQ/6h5xoi3V0FYNO96RZO7AGxilUMiE4hsmjrrbUOEMK5BzqikQGKa2UKAud+bMD7fJEjy4m2syA8lV9JgmlbNOC4yXVYUu0BOqCu8pE2vZIXHUINestlCCgT6zZWlfvEsChJF3yzMRnibdkda5Va523GtjFwVC52imVoeYlluhVckuYU8Fdzvj137xXQ/v6c1A6jVLdHP+i3auPum/ZRd0QdDSpUuZNWvWOX1nTFo4P941irKGdsL8tXQP82dkahgnq1u4NCOa4UkGcsqbKKxqZkxaOFqVkm/2l2Dw0zK5dyT/2V7E0EQDN43qTniAlvpWC3WtJrRqJdHBei7pG83nu052JE21KiUvz83kotQwNue5aEbRQXpqW4zUthgJ9dMxd2A8L67x7J4zpHuox9+EJnprVFTlUjLgb+gjhhNbvp6K8FFs0Y1G6xcIEUlw41p5SdRa4TqHpcHlH8rL3FIjjJSWKshfg/XgAFRr/yFOYeITXTeCCEmUaA1kv/hh4hAi0h04tF2cwp4PxcEpFNJCrDzLxYZInynR6Y/3CBQy8SmJ9C3twlOOHShFNzH9JIlWvB2mPC/Hd0IcVrNAQhc/ITKuu96ViLT/NUIt6+xwcn8SPLtoizR5jh8q200JUkSz7nGXUp9vmDi3GW/JccY+JCX4IQkw+0M5/8SRnpogFjNED4T6Iomol94mE6FSJdH4wOskGVxbKI43LFUcT3GBdxcgpyBZRLrkJ6pyZIVQvANmviNYtlPN0W4Fm0n48+E9JYIf97DQ/sqyZCKoypH9q3Jl8tn4DOz+wPEgZjryAS/IyshZ1Zp2iUxw7iuAkXeeGSP/tWZq9k50g7ccQCf7xXc9LEWw580vSoDT9wq55uejyfY52G/xSf8Nu6AduZN2c66WYPAjweBKAKmUKtKiAulm8KOouoX0mEDCfLU88N1hDp9uYGSKAR+NitP1bZyqbaO6qYyXembyzs/5HDntilofuKQHY1LDuWOvS8jKZLXx5oY8FgxPJDHMjz1FtWTGBZMS4c9D3x3h/QUDGdsjgssGxFLdbOTzXSfx06l54JKe9O/WheTqlYul7LyhRMq5DSkENRTzvW4KW+yjKcpuZ25PDfNjVSJmdXCxSKf2nCaYsEotRTg+IVIgkrVYHG7aJaiaHB1+jI0CXTRWSKJy078kIvULEwbGpucdzv5JaQ/WUiXOadzD4sSsFug7VxKNCpU44lH3iFOxmqXd2/K75LdqCiRBN+sd+R2fEImenU7yxCbIWiLsjK4KJ4xN8tuj7hMeelOZS+fb3XT+0HOGMEXaHE4tPE2cf+9ZgkU3nJLI31lIpNRIE+fljgrNsoMiETBNDVfoAAAgAElEQVT1JdjxtkT0PacLPFK6VwpO+v9JJg3nasZmlSRw/FBJWnbGZP0j5Vo69w+MkcnDiRlH9YU/rxO5W52/OF1js0gL5L/gOs74Rx0SujGSAK1zSA9X50rFaHC8FLwkj5f71lIlE7WpWc5ZpfWUsI0dCNcuFyimvRGG3nz2lm6/1nzDxMnueMtzu3vXpi7sV73rcYNEC8fULJWz/wP9M3+rT/qj7YJ25Ofb9BoVPaMDKa1v48jpBrY65GpXOhokj+sRwXOX92V/cT1Gs9XDiQO8tSGf1+b19zruyZpWimtbqWg0MjE9grgQXw6easBktXGqVuCAmGAf/jEtnRtHd0etVHYpMYBCIS/iTRvkxS89CIuvJAy4MSSJKZl3weBEYhv2ovDtBgvnCQYcM8DR6HiXFM1YjZJo27dQYIKIdHGUNqtUIYJUFE58WpT7Zr0j8EJQLOz/XKL6CY/Cdze7xJtMzaLBMu5hmSBWPyRRJUh03VQmpehKlbfeR3OFRI09p4m2t9OJt1QLla6hWJyjIcWzWjMwRvByfSA0lsO314sTvNixmnAmOBUKSJsq0MIXDphB4ysRXDdHZO5nkIIW9+NHZ8qEedH9MuFkL5frdXy5rIoOfSn3YehfhOHi7NjT2EWXmMbTriIqkMYalUdlTPM+h/VPiURsdZ5UPVqMElEXbZFOP3GDITBOzqvskPQxdbdtr8FV38g16ZwQLdktVaognP2CDUJ72/icFG9pfGDm28JKcppKLYyebiMkKj9P9DiUShh8o+D1x5fJfZjwT1dD6N9rGp/fFoXXFUlepXg7dB8r1+t/IhH6B9kF7cjPRoD/rdbYZuaJ5UfpGRXopRm0MaeSfvFBbMqppG+cd9a51WwlwMe7ifOYtHByyhqZ0S+WhjYzC7cXkREnQjndQl0ysmqVkrggnWdXcovJVfXobN9VeRRUeljnajumrCskbtNfYfor0GOyOKTEUUKzOrlVysqjMuSBDUsVKGLUPY5jloJvOKy8Vxy2SisvV3O5OK2AaIl2vl7gGldQnDf7xWaB8N6w/z8uJw4izNRjmkS/1bkujD5mgLAIbDZJvq36m9ADnWZsEsdYtFUc2KRnhG3S3iB6HSkT4dubBG7qO9d1s3a8KXBIY6mMKbKXKC9+d6Pr2OZWwdsv+ZdcF7UWWt0q7GIHCbd7/eOCwUdlwKSnBILRhzjofjtlUuoz2yWf21Ij16ahxPPa+IXJ5Ghqlih80TxXgi4gRjS6P5vpOgeFEg58Lni/02a8CUd/kD6f4x8VJoszeW1sEmio/qSwVFImAnZha5havBN8bQ1w40+Yqk+iDYmRhHLnCNbkSBZWHBYuttoHMufLpKL0fs5/tYUmibZMwz/kvgZ38xxf/SkJOmoKHHzwwX/Iu95hLTWw9HZXrubwN0K3nPGGJ3XzV9gfOs7fYRe0I8/Pz8dg+H0E/s52orqF1UcrqGoyMiMzhh8Oul6kIYmhlDe209BmJiZIj69WRavJlTSc3CuS9OhAXpmbyZMrjlHfamZ8z3BmD4ilvLGdb/aVsKOwhjvGp7BodzE3jurumhAqjkmEV7JHnFKPSyQi3fqaYNBxg8UR1heLUt7Qv0giESSSjezj0g+pOCrFGBpfVyKvZK9EYomjJPoZ9w+JyJPGSKRXWyACS/s/kchyzSPCtc5ZKRS7zlKrljbv5gcKBfgEetPAQLDjTc8Kx3zwn8VBGBslelSqpUij1yyBg6L6SqTqH+HiR1vahQ438HqJ+P0jRHsbgIBOjYWrRQ4g7RJJ5B5fAcEOtTl3fY3qXIGWzG1S4dprlhTwxA6UhO+y213OtjwLdjdD33nSjq30gHCUCzbKPRl4g6xCDn/tcLLPy2+pNDIB7XwXji2V++gT7MmyaCoVyqUuUK6nQiHn09Qpst/yihT8LL1V7u+Qm0WMCiSK3P2eTDhjH5RiF4VCJmuLxbO8XqGQlUhIIgdyKxia2km4zNwuK4EtL8s4e05zUUwPfAbXr/ptypXupvXxaMTRYc2Vwut3CqwBXPw4Jf5jf/+7XndS7klgjGdv0upclxN32pFvYORdsio7B/sjfNL5sAvakR86dIihQ4ee12OarZLc2V9cT1yI0AeLalroExPEoMRgSuraGZUSRnFNKy/MzuDL3SfJq2hhat8orhuRSJCPlssHxDEsyUCb2UpMkA+bDxyntF5JkI+af1/Vn+ggH6b0iSI53A+9Ri0P2OeXuxJ5xTslWVWVI9FdXZE4GBDGxcy3oGCzODWVRpbOPz8vn/e5QnRJTM3e1ZqttUItjOwrjQF6zYAf73NBENH9hLGxR8qAO7bb7XgJTGV9JUnKlfeL41copICkqUyScdnLPffX+ksE/9MD0oTh1C6XprfNKvS5sQ9JBFRxRJguWj9ZCThpd0HxEqFXHIEmg0xGW191NC0OFHzd6agVSsF0f7hdonK1TiLsrW7l62mTBU5oq4PTB+T7s94RZ1p7wpvSVlsoOYPVDwtUMvKv8vepneL8ZrwJBZtg/6cYL30XnY+vTB61RXDs73IMv/CuqXZNZY62dg0y9q7ExIwN8ru9L5Pkq3+UTNbpl8oKYs3DkDhGJmGnbXha2rPp/OU58jEILObgUnf5Dp3aDV/Mcf1ddlAm9eIdMgEf++H3O/IzWcVRTycOsOk5dON7AN6w5a8yq0Um8+V3C9QYnAhzPpJnC1xiYZ2ts/7Nr7A/wiedD7ugHfkfYd3D/OgVE8ix0kaWHSpFo1IwMiWMgqomQvw03L/EpZOgUEgF5ktX9CMqUI/STaUwJtiF0508tI2HzlbtVXnMu+Bk338kmlLrBd92Wnu9RGHBCZA2WzjR7g/+kSVS2GGzyPK387OoQKJprY+II7nLuJYdlOjbiSk5WSuFmyQSdJZ9g6PSM0Wq4+w2SQye2u2iG9YWykupUAp7wImlD7pBWAWdHT3IxJUwUoqDnBrn8cMcfSIfFQhjrVsXe99QiZqOLpW/L3lBOOzVeTL2vR+7NLIVSolWtf4yySVPkEKUVX+Xz/pfI07SxyB8cH0XDYM1vnKeo+6WSTEsHWxmuVZlhySCHX47XPIvFn27SqhodrtE7U4r3iGRfGfmRvIE1322WWUSU6o9tdf7zJZVlLNPZlgq3LEHdMGweL50OCra4j3urK+kNV3mfJnQnC3ZujJTuzjqzpa3RvDywo1dN8U+X+bepchplnY0ii7487/Wqo7Dtze4HHN9kUByN66RyNyQJhK77sn0pHFdd2P6X2oXtCMfMWLEeT+mwV/Hs5f14ctdxWSVNNA3LoiYIB82ZFcwPNmzW7jdDuUN7YT76zyc+DmPs7NUasc2lTAGOlvlMYFW9n7UdcXhya2y7B54vcAwTguMFWepDxIsOrS7JHWUaqncaywVJ6fWw7A75OUFcfb56yVReuwHiVx7TnMIb/mIM7OaxQFWHncUDc12NFh2YPBRGTDnE3lZ9n4kS1ancJbTgrs5eOrhct7lhyVCNaTAvEWSzHQ3lQ7CekrBz8FF8ru+BmmUsPYxF24Nojmy+wNJxvpHCLfcOSnYbbJ6Gf+oSAukT5c+mINukPPyCRaOdq8ZsqIYdL1E640l4jjDUoX10v0i4V9nXsmgQYMckfhSFzwEEv03V4rW976P5RpOeFwgr9kfwfa3hJYYP1SSoNtek/17TPl/7V13eFRl9n7vTDLpvfdCEggBAqH3XgQUEFiqGhFx1V3FXbGhrl3Xsrp2EVdUEJAiIEjv0jsk1BAgDUJ673N/f7xzuXdKQjCBmfCb93nyQCYzd87cO/d853vPe87ha4vSudB6RJCCkhJyfZ6hwsdUJz0XPyD3IpB9kvkI/3Y3aA2972beRY68M/VV1jjruHYVi7KagswjTJoXZ3BRC+9PSg6gkshw4HbMSNj4NMGpFl4xjq4LLpEucvZlG+GJC0iLXdwKtB7NGoeGFrx6cDt8UnPAoh25i8stjAu7BXQM8YC2ToSNSsAfKXkI83TEB3+JhygCrvY2KK5kRGKjEjCrXyRsTQyWuCU7fdvyxhQEOpy6av7uGsCb59hP+txu65GcbJ57ntH31RP6x/MIZ1n6gBcplbu8mwk2R0/qhTtMpnM5/B2fr9ZQr7zjPXLtvWfz/T3CmFR18afyxD2M7/3HR3R8XR6hs804yOPY2AHjvwdOLubcSQn3fUb6KP8Sk6f5qfycqdtlx+MRwShcUPP/v88hfw0w2Xr/PH29dXAXKnjWPk1nH9Gfn3fHu3SmQ/7FJGlRBp1k1jEqEtL3cVE4OM/4OkiRZnEmz3dgAimr8lwgfiqg0tCROniy7YCkvjm/gbx6eF+gzUigrgYxThXAgZ+42kf0Iw217zNG8tpaOrAef+UiKkkN243X7YhsdMnXAhYelWTxfJblsEukozeTtFJitKaSDr3NfbQ9aYWiCMyeic/vh8nn2t6NPXcC4+XvZm0VP+vJpVT9KFsKqNTU6l/cATz0Knn+P4urJ4EFo+RjX9jERG+8Lt/h1Ypqop3vA9dOkEbq8ghQ2oTkqrKoSoKjp1xpC3ABGfQy0P85/YX3FnG7fFJT0eTuh38Gje1+OG/evFtqmnWryC+rRmllDbyc7W7MtTx3rRiHLxegqlaLzmEeaB/k1mA0btLO4qtMapZkA36xgE8se2TkXwRW/00uwghKIBfp1wHY/i5QnA5EDaMDKEyj8sLJl84rXze4IagzX7P/K1ZCVpXw5izJkgt0xs0DVj6qb6RrECspMw6y0CZmBG8gRy8uLM6+7C636H7ZvoEvyVPYJQQmkDfeNJeRevxUPn/r63SogpqJPVsHFiZdO0mbKgo5jGDMV6QwfjFojhk1lJz27zp55OBXgUu7WX5dnEV6JWowHXfWMfLnxVmMHhdPlikk/w7crRz/mY5dQqfpXKTyLsqTfCStu4SEh7goDnqFDZoMMfJDOtXQHsCiiUCvJ2lDcRZfW1VKB/3HfziMQeJoJZTlksY6vog1Ax2nkOK5uJWLklcrnr/D8xmlx09lb5Yd7wHJKxg1D3gZsNXIMkr3cH43DJPVXR8FRn0ofzcL04HPOnEH4ujFCtWSq1wI2ozmdW0OCeKh+czLKOEZCTy6Td+x1lTq9OCegErVtHu9qpTf6cPfcWGwdeTiGjX4z3+OenC7fdLNcDu7H7ZYeDpp4Omkvzq39ndFa/9b33LdQEUhv1RuOpoj4wCdpr0Hp6QrK+kyjwKdHuQNNfoToOwao7AlU+SIq91EXcR7mRFxWQ4jSCcfbidzL3CbKkVjgQmM2gxRnMmf6jLqykN7sAn/uK/ZIS47mY5Psk+lNs1n5qXQgXuEMxFbVyUnYq/sYZVn+gEuQGd/A04s0X/9kQV02IbIPMzH+/6T2nLvNuSed31IRzDoZVInbUYymnbyofMrucZKTrUtqRi/OHL5vf5GB1eaTcessiGXWlMBFFQxKjbEuXVAt7+apsIALnhhvbiV7/csI3C/OCZ/939BiWf78UBlKaNqQ5xYwgUQoMLoxM/AjE2kcnxjgYXj6dz8O3BhcPQEDn9PhQXARXvj86zKPfojH/MIJ4VmiOun9ekGjROdft4FVuBueY2Odfx3zZzY1AU9roFUZ6ltABH8UcLWvulVpBLsnJnEd/Jm0FNTzuvt21a/COouhkU78sjIm4+IsgRERkYygso4zAg0vA8jYqnfxKll7CxXlG784uJMJtGGvgFUFLMlp7JEP2kZpXO552Qpmq0jBwun7afzGfYW26KKIh2aqQZQAfGkPI79xOTiwW/pkHd9QEqlOINO3NaRN4K2Tn9Yr4TY+4A6nS782E+kSVQ2fL4o0lGP+4YLmgAmD4/9JL/e3sV0IUZ4X7a4PfQd1TF7PpEpnfxUOp4hr3M7PuQ1RuvdZtFBbXiedMagl1n1mH2Kzn7YW7wuLn5c6Da9rJv/qWJEbwjXIB7v1DI6bGWS2TeWZf2VxVRIxE/hohDUhcoRCed/Z1dJz3CD65zFc61EZZGux3wbnvPqUu4w7FwZ1cdPNS4MAoDKEgYAZdmknGwdgdNr9J/T6QFApZbvIUdPYOT73ElIFJN/B+4Ws5P5nfOMvPVBwWW5vPYuOnojpCsTi+3GUVpbXcrzGjMccKx/wWjyvX71hLybA9jwzcmHu9tmhKX6JIt25P36NUPJ8B1A/15dgW2vUUVh68B+HobT2Y8vJA95SsEtCwITP6KWuujpvwJ7P6ZqwTUIgEj1h1ojO3GAN/3O96kWqSwiJzr4XzoZngMLUAa+xJ1BVQmTdf1fAH55gM8puMLRYEEJjNIqi9nL4/jPwKC55IbL8wC1AymCbW/qJuHcA3Saxqg8pCsTdjmn+Xn3fqrTiyeyrF+iOgI7yZJHlZp0TlUJnfChb+n8vaNJQ9SUAX2fofOXnLiEumouUM4BwNpnGAl3TgR+nshz0PtpOsqKAmq+/dvTedrYMwF4cqnswEQtP49XtDz/UmXD4cNb/sX3GvoGueIre3ksRy86h7DeXOCqislnG47Mq63S77MuX+yGu/L5xNIB+sWRzmo7lu/r34FTjy7KTdOgUnP8na0DC6FqK9mW9+gPdKpdZtBOGNxDEQM40Dn3PL8znpEspT/wFc9JcHdg7BeNG5hdVQqcW8/vRm0F0Otpnj//DlwwfrhX3t0VZwIrHmVfIBdfk4dr8r2uPD8Sjv7A76upwOZPwlJ90p1vVnALWLBggblNaBRqrxykEwfohGxMJFOunuAWOvY+RoSugVQzOLjzd20d+ctR/2HWf9cHdKitBvGYhihK400/8kO+1iuGO4LzG8mlukcw6rznfVIL29+hI4sayq34tVO6BllZVDuUZQMT/8cbdNCrwAOr6XiuHud7jPuWXRCXTGUvmPW6KNinLemLgXP5c2qFvuQx6xibPfV+Bhj6Nj9j0nIgejg//4QFdLw1FaSPamvIE9ub6OdclgPs+ZjO+vIfdJhShamNHT/3hO9pT/KvpA7821NyqKxEBbgwDnhBd44+YLOsA9/QiQPcEajt6CzPb9S1+xV5zvMu8nPYOChmhCpgqirSNYC0kRIOHvKwZNcASghLsllUVZxJ6eT6OVzIE3QTlfzbc+EqucrPHxhPiidpORPMnROpPCqnHFTvHlKpWAHcfgIXIaltsuRwMw5wwW2Mvjr9ALByJmmm0uukjM6v598qi4zb8RZeZiVxPWjyvW6qC6J3NJP5zQhL9UkWHZFbPLRaoOAy7Ep0MjhBYFbcxo43n+QUAN48KVsZLQ94kRF32l7gwkZuj9P2cVuf/Cu5dYDOYv9XHNNlCN9Y0i6nljPqHPMFnXPcOGrPw/oy+u/+GBOn7ScwCnYNAo58x3FnuSnAld1cAAA63KFvcjdx/Gd5TufRH4GoIdw9SOqa6jJuZcd8yQgvaQUdUL6JyfaVRRwAEeFG3jq0J+VhaltSQtLuxSOc779+Dot7drwrHyN6KOWMalsWDpVmcyFxCaBTE9RA3znUD0uR96WdpGEENXcka58mbSEIsjPc/jYLbnzb6Zdrl+eTsqgul3MOAfGkf8pydAVOc+mI18+RX6dxYv5BiZLrbN0riuSkMw7T0XWcIncjBLjAhvfl9ctQzF89v4G5hxHvsUhoz8e8/h2ny3r44ix9KWpjRAzKugEJ535nTYE0oLo+nNtg/Nih7/g9czYxRMLRSz/ZaQqiyEWyOIM5Fq8o00GRKXjH8LstXStbRy5qNiaCoLsQFu3IHR0db/4kc6GqlHTJtSSoHDy4TfWM5I2Qf4lyupNL6fhaDeKNeXE7C2ykuYd9/kGO1COcE3eKM/VLrSVU5OuqHP/DyNU9lNPZpXmStZVssNU5kRRHdRmjta6PsJ+Hm477TVoJdJ3BbWhgAgCt7MQBJgMv7WIStv146tWlKsWULVyAlKirBnLPUk+edYyfu829wLEf9Z9n6wAEdaRjnryECozgbpSmKSmogstAUQa0A+ZCVV1KLv/aCSobqkq4YPT9JxO87qFcMHs/RVVHWQ4/t2Exy+nVVOTETeCCmHUUgEBqSNpFCSpK2NqOYWvdzCM8P96teE2jh/D3qlIWnvBFXFS8YqifP7eOu6y4cYyaJWi1wJnVdPaSc+3+OCPZWv2hJgBIMe352PjxjEPU6BdcZnGR2pbJaN9YnkslFRUx4EaxS4P3kE+s8WOhvQBNI6gIU6PTPMIo4fRto694UmuosGqgd7iHhwe/D8sS5WTlyA+5y7iZM68p53er/QQ6cFHLXVH6Qd57zQhL9UkW7cinT693frP5ce0k6QW/OHKbnRMZ7UlDf6+f5nbezpU9NCS6wUbRuS1lM7lFjwigIJVRi38H40IaBw+qIMZ+Q8VKdhIrIbs8zG1/5hFGc+v+Ib8maQWTPdeSWLTiEU4nu/sjcpkOHnydIdL2AH2fo/PpPAOoyGP0GtqTDjVQJ2GTlCdO/lTbCAILUapLGD2nbOFnHzhX18rVBRj0Lyo3Ti2lU1MW9EgozoBq+DsclJx5mKoVgJRTUYb+2LugLnS8nRNJxZhS2UgNvGrL5ZmjAHBQp61P3QFEDuD5qSlhZBs7ho57w/O6OaCjuWuqqQCmLQMyjlIW+essRu6CivmCvs+y4ZUSeed5rZQR8sGvSXfVVrAHj7MvFRcAnV3kAOPagdAecgn9ud/5Xen/PLs6jvuayc6UzVRIxd57I2nZ4D0U2oN5D4kScQ2iLLExUWz0MM4hlYah2NixhbDahovvlX289nVVzMNc3stFNmaESbXKxCHdgXn96JQBPnfdP3iNA9obPV8Pgpo7JcOum4NeMf38JsBSfZJFO/Lly5djwoQJN3+iOSD1g85OZmLqlI73lVqx1lRSpaKsxAvpLk+mAUizOHkDu/7NaERlA9z3BR2IVPnWbgK5TLdg3sQbnqdzcg3mF7fvs4y+TbVWPbOGCpDt73DL3mogt6AaRzrZgHg6BSXC+wHrZusWIRfac3IJOeKI/jo+uxzY/DIAgdv72FHkv+1d+LfY+0iNFKYzKdfmXr5nbQUQM5TtTVM2U1op0UgSgruhbMencPIOAjyjGX1e3Ep+9fgi/edmHiZ1lHOO0j/vKDoXpeon9j7uSrKO6fdWqasG0g+z98qe/3KB/MsiFt0IoKZ+4g+kcZx1TtbWng4suBu029+FSnJiopZO9sTPlAYCusHR6bwuhj1dRJHUwarH+ff4qeyZXl3GqL7DFODCFrmkPGIg6wcmLeSuR6VrRyAlJb1aMUncZ7ZRQrXBe8gtiJ8/9xwDDa+oxrd29WsLPLxBnsMqdd+sKgE2zGVhlnJq1KCXgWUP8TVhPY0Ol5VyAoHKojjpvJZcvbkjt7Hj57+8U14wNU4Nzx/9k7BUn2TRjjw/P//mTzIXlDzgjndZ8uvoyQTeSp1KY9BcyuSunWJUF6LjTv3aUS0RP5lJI6lUXlsLbHgOGPUxVRGVhQBUQF0lFw6vGCYixVrSC27BvPE7PSzLv5TwjOTCMfBFJvKO/cStZvxkUg4d/kJHd0YnXQvvx+hSSgz2/BuLZqRFJT+VDjVqKBedmgqWobvoqko3v8obuesjVNUUpTNqPPajHGG2HUOFzYYX6NjjJ3P3YONA5YmNPZx8QungBDWjqnbjeWOammcp9X7ROFGrPEQ3Sq66TDf0+QJ3H4ZTegCgsoBySddAXp9lD1KS2XEalT0Xt7K/S9v75HLumgogdSdU/u04lKKikM7s5FLulCoKSVedWMxFrOujMo8vwdaRu4viLC5U2moqLPzb06n6dWBJeVkOHez5TcD8wcyFDH0d6PgAh08bwoQq5qb3kKOHMaffWPi0Nu5wWFHIAeCGUA7HNuHIcyrVCHT0kqdMAQxYTFE4phDaC0hcz52gxpntk2+xs2FjYKk+yaIduUUjIJ7RctJyOphzGygvtHVmW1MbO968nq3IYVaXUO1RVwXETyM/fnoNEztKVBYxSRnag5xwzjmZAvFty2Y/ygHN/h2A4e/xuL5tSekAdGxx48ml7vi3PKHm4jY6o/ip5OTbjKQjrK0E3EKB3/4mH9vWwXh48fkNQKshrEyUkmXFGdza9nySPLJXFG3u9Xfy7kqa4PRqSuPUtuSo244jf5qdxN/94kgNZCexzD6gA9BlJheiyIH6UZ5rEM9XXbVOQ2+j63V9kE7v7Fpq2jtNo+NMWqH/WSIH8Fy4+FN5ET2U0fvez7gApWzhuDr3EDr7k0v5/hEDyE2v1rXB9YpiUtU1iFRNxiG5iOf4Qi5GB+dxIXQNZOJSSpCG9eTCWZoj00b+HYDhb/N6bH2DOQWA12j980xUm3CGfxq11bwezTG02NGTAcG1Ewwk8lKYJ5IoRScfky87npqD+An/43DqykJ+90b/lzvIxsBGw3PSnOelBcGiHfmkSZPMbUL9cPJmI6OY4XS4opY3dkg3OjXfWCZ5TixmlaFyqsvRBdQi56eS480wiGBCutMpCGqqRyS0HknqQIlrJ4GsI4w4uz9GpyDWUbLoFkQOs9N03mDXkhjxRfRlRO3gQf6+2yyqTexcSBMl/8pjm4qA1XaMBpULUGgv7iyOLODvoz/hruPKXtPzG3POsd+LoxeTd6sel/9WnEmFjESjFFxi4VP7iboOh+3pYH1j2ZbX2Y85gKDOQEkmnXKvp+RcRWURn2vvTqe+9zN+rh5PMFJe/rC8yAV24rk4OI+Kl6DOLNmvrWISTvosGYdZMNRqEHB2HZ1V8iq2Rtj3uX6Oo7qMxUyDXpE1866BXDgAXis7V34H/Ntzwbl2ksdL28cy8wFzgR1vy8csvNxoh9XgPVScRftPLAYCOgFdEvUTtUpUFnGBTD9ABx3ay7joCWAAMWgu7c8+xe+ed2suot6t6+3jMnbsWMDNDXhsF+1y8jY9DMPMsFSfZFlnyQBJSSacgCXBwYOc9a4PmESsKGDy0TeWOuerJxmlFlw2fm3qdjpa10D9dpqRg+gE0/YalyfZHREAACAASURBVIqrVPLEGCVqyvm3k7/oZnVG8hjleXyfQ9+y+tCrFXn83f8BVj/OxlWdH+J22DWIPHGHKaRUvGO4EPkbbE+7zgSqymXu3zWQDlnZG10Q6NRzzpre3gZ0ZOfAyiLjHiGA/vYaYOLOpw0/h1swE7vZyXSKZ9Zwkdr2JqWam//FXUf0ML7WMxK4eorXKPs0lTfhfdmf/MBXshMHuO138NA1ugqgHSob0wOFr+zlOZaQcZCctsqG51mJ2koutPu/kK/3AJ3iyMmP79njCV6DjtO4yGUdo7Z872ekoEZ+KGvWXXRl51otZ31mJ5umjtDAPVRXw06Mvz/L3dPh+cCPY3g8Q4gidyqLJvA8rnqc06RM5WXKcoENL1LbnrqD9p9Zw880bZlp56+00yOMi5R3tMU5ccByfZJFR+TJycno3bt38x+4tlIu6vCIABxMFJ80BtWlctZeCW0tI7oH1zDqNuV8PSJIn9g6cRstjS3TuLACs7aSBStSyTxAJUCbUYyiJDj50DGkH2LVZF4KqYb1z1FiFzGAlEHqTh5r25vya4vSgTPrgIRpbDtaU07ZYWhPRroOHkDPJ8gzl14D/DsyoViWS35YGvCbn6qvypBaEeScJQ+fdUyudowaSo7fM5zvE9RZXz3jFc1oVePMRaf1CC6OJxaTqslOBsL6kEIRVFwolHpzgI5j0MukYlxDgO+H87P1fRZYOo3P6TiVSWRDlOfxswd15vnuN8d0awVAn4rwimKuI2YEd2ipO+Qe7Z6tSHtNXkyqKC+Vu6KQXwFBAxz/kbSNBL92vJZ/fKT7PKuBaN00JDtXRs1VJcx5bH2D9E54P2D0R0ZURL33UHEWI3/lsI7yPFJzXgZl6IVXZMmshGsndJOqAvUfzz1vXJV7YZNu1muY6fPYkJ0WBku1s0mOXBAEFYBHAbwJYJAoipa5XClReh3Y/TElYKKWfVHu/e+fazLvGkglx6Wd8mMqG3LIvnG8MdbOZoMl9zBZbmfnSqpg9ZPcWnecRsex8SU+LiWGDn7LLoBJK+ignf35XJdAbvl9dFtVtR2jtr3/pfMrTGOCriyH0jeAjsQ3zvgzhPVkdCVV813aSc5aFMnxb32DztqnDSV6iybI/UpGfczFobKYeYEzq3X8tEJiuf0dtnRNeJDR8clfWGIPsMf4iHfIu2ccJs0U1JXXpecTdODHfqJzPPANE6h9/qnb2vcAyvLpdE1RQG4hnKFZW8GqyMyj+vx6+kGOwVNSVwDzGTEjSFG1G8/PqXEGQnpSiSGhzSgeAyCd0PVRzv8M7cnFeMhrXDxs7HntruwF4sYw+b1kKncCgoqVqMq2DQCjfxuNPBfU2Z/Rfu/Z/J7Yu3Jh3qDQ9V/eRfXNqE9uLh/MOceFUVtL6qOmQp4aZSoKrqsxVt4ApuWeVpgFTY3I4wEcAFDeDLYYoX///s1/0PQDwAFFBdzlP4AjP7Ah061s5QrTqWzo/wK0PrFQHfwaCOkFDHyemvBkuUwaf3xC/trOhZV4Lv5Ug0jFMIfmM8pqP4GyNAnleSx97vZXqlUyDpJ39AhnxWH2SS4Qez+TByVfPcG2pFGDZYcJMNIy1O+6BFDOZliSfewnLjZBuik+tZU83u7/yAU3Ef3oPJQJxL5zONC3KIPqkR1vk18+v5GO/vc5+gOdizNYkp62n1Fo1nHy3/1foNOWcGUPP/+Wf3Hws5M3nXBdDQAH/SQvwPN79QQXljOreezW9+hH4HkpVNBED2d1ra0D6Y7grpyNqaR8ghKAXrMpZ3P05GLtFsIkYex9pKR2f0DKKHU7r/O59TxvEk00aSE/66rHZTpH1NJuU1WYUiGW2paKmi2vAyODZfVM7jnj15xeQ+22Iko2uoeKMoGfJ8ny2bR9/L6E9eLflHSRBPdQdjKU+scD/Iy+bYyf6xVlvMuKHk4FytY3uINtO5YzRW3lBb/ee70og4lgZ9/GK1huI26LT2oGNMmRi6J4DACE5sh2m4CNzW1gfgwTiwDVDX3/2fjOb5lHOSW99DogCBB7PAU8eZgR94XNTNSF9ZKjGG2t3PRqyOssipGGGgsqUijpBxi1hfRgw6Ztb9JRaesYfS5PlJ2CjR0lfCcWMQFqOO3+3Lp6WsUeYoOrg9/wdzsX01NSVLbcKajt5LFwgiBTPAA/n2Gv8n2fMpq3sQe2vcGko2ckb+rsJP1BzjcgMgI+tYxJzrj7OeRYCW0dqR17NyaAXQNIT2x+hQ5i+Lt0SJd3M6IO763fkTA7ic4jbhydtkR17f8KmPgje347+XKBLEiV56feOG9HgVKdfHD9c3y9WsPhDyoNcwwRA7hD2vEO0P9FRubJK/i83rP5e+l1Y5om8zCvoVLP79mKDrXfHDq7vZ+RmvBUUB4uBpQGQLWLQYOoG/dQeT4Xt/I82YlLOLeO1am+sXTahrCx4/u7h3L34Neeum3D5lq11Vy0IwdwQbx+mgFAeF/gu2Hy/XDke2DKEi6uhnYqkbKVRVdlurbN4+fz2GbEbfFJzYCbWiUIwkYAJkTKeFUUxTUmHq/vOLMAzAIAf39/zJvHCS7dunWDj48P1q0j7xsaGophw4Zh/vz5yM3NRWBgIBITE7Fq1Spcv85M//jx45GSkoITJyhr69WrF1xcXLBx40YAbDXZr1+/Gw1uHB0dMX36dKxcuRKDvINg6K6rArri6tU81Il52Lp1KwAgJiYGXbt2xaJFVE+4urpi8uTJ2L9zM7qeexfq0uukFTpMgtrWFlWl+dAcnA/BP46cY+pOqlWUSTJBRUdTWSRX07Ubzw6H107x94LLrIQb9jYXmJh7uPgoE4C1VUzoBXcxLRlT2ZAOMERdDVCdw5vSNYRadrdg474wCQ+yB0pxJqPUvZ9yG63kU03x/rVVdBiugXT+uz4grZF3gTRPh0n6SVGNEyNOn1hg2Dvy8GFbExpptYbv2W8O27jWVsnNxtY/BwQmQNvzKRT5doPHzyOMX+8SwHM28GVSRio1q2UPfIXkjq8h2M0LS+fPx+D2gWilLCjSQXQNhrBihvy566qBjXOZGE5azh3Dun9A6x2LbNcOuKzqhp795+DCxcs4kVaI/EUrMXPqOAiekRCU3RFTtqJk0ioIXvFwTtsCMbgLBHt3KnF8Y4Gacmj7zsGB0iAk/+97zJw5E2vXroWfoxbx4QOhuayjizTOyO30N2Scuah3PxUXF+MfMyYAG16C6vw6Bg1GEJCv9sHyX3cA2FHv/bT8ajjC4t5Gea2AoUGdsXz58hu66kmTJqE28xi8fvu7ruDJh0np40tQ7RQMjQE1o935AVQR/bBk5W8oLi5Gbm4unn76aRw6dAjnz59Hv3ZBaLNjlrxrLcsBfnkQVYlb8MOaHQAAT09PTJgwAQsXLkR5OYOMxMRE7Nq1C6mpPMfDhw9HSUkJ9u5lS+L4+HhERUVhxQruJH19fTF27FgsWLAA1dW8B2bOnIlNmzYhLY27olGjRiEnJwcHDx5Ebm4uhg8fjuDgYKxezdmngYGBGD16NObPnw+tVguVSnXjOmVlMSE8ZswYZGRk4MgR7lQa8nsAoNFoTPq9+tAsE4IEQbgMYHRjOXKzTggqTAfWPAWk6tpeugYxm+5ngj82hYLLwH/j6cQHv8pIuyidTq7fHDrpw7pEYNQQyhHPraf8rf1EOt6SLGD/13xO/+dkGkHjzAo/Jx8grB/1tFVF5I+PL9S3IyCeUaZ3DLesyq12lxlMKh5byGgLYJfChOlyf5Y29zJaqiwirZO0klRAWE8dHbKJMyJdAsjLh/WhFn7tbC4q/V+gCkOplvCK4o5j/XNM4P76GI/Vbjzlak4+TICeXsVkb+eH+PfdHwNdH6YDP/gN+WflEGY7Vy5sEFmkU1nMaE4UeaOnbGFE3nacrumWbjjFDbtaMdIX1MDO93Stg+to+5DXKCOVUF0GrHyMQzEkOHiwOdiSKcbfh37Pso2AVzQrQV0DWWRTH9IPAEum8dqrNeyx02k6I++6WspAvxsu6/fVGiBxnenhD2W5DABKs1nIVHyNO7GADjeeMm/ePMzq4S5PjOo6k4lYacIQwGrde95j5N0UnP3d9Dka/x2w4hH9x/zaAzPW39hBGN3rl3axFa4hEtcxr2UmWCcE/QnExDSyGOBW4B4CTJjPhE9tFbeHjS1LBlgpGdKT/OC+z+WEVEUBnc+AF+kYdr5PB5N5lE4oV5esvLCB3QR7PqlruqTbnnpHc1jAofl09CXZLIxx8afzN3TkHSbTlnJdQ63c84x8/duzfH/1E4zYh75B53J5D6PXkO7cotu5siilrgYI7EyVx+H5pIYSEtmQP6IPW8AKKipX8s7xswgCXz/mK/Lg18/wuG3H0KF0/ytQVQFM+pkOo6qIXQ+ripkIbDWIjvfnv7CnSZ+nWenZ+SE6uppyOuScM7zRw/syCv/t70yI+sZxsZDQ43EuSAkPkoIK6crHpOZXnhFcwBISZU4coNSz7Vj986px4jnzjuLi5t+ePHhNhb6CSHquVpdozbvA89KQEwd4nmbtYEDh6EkaRRqxprZhUnnGes4Mrasm39xQhaK2DvjtaTlBvvdTYMZGltCDESgKFY3YDv+PVFJVKXM8saN5PZrqxAF+zyQqToKzjrJS2eg3NOszW48GMrrXHb2NX6PW1FtQdKdwW3xSM6BJEbkgCB4AngTwTwA/AfhZFMX9N3tdYyPysrIyODmZ2GabG1nH6KSVPKyEez5km9h2E5icKcul+qTwCmdYXj3OBUBCaE86Xr+25JUlPrr747wxynIZUVYWU0ddU87pNIVpLNyxsefOoCSbzvLCRnLJ7ScyCZhznq9rM4pO7cJGcsXKRChAe0O6A1tepQ2DXgEgUD1xfgN3LlGDGT1fPcbX9H+e2/+iTEaFLgF8r/xUHYe/mDsdqcDINYiyP8NJOb2fZt6gc6I8Ng7gAltTzoZZOecoMxz4Ev9Vfm9VNuRcRQHIOsyktY9uh1WUxgh8lW5SzL2f0jGKWu4g7F25oFeV0rFKVJUokorZ+W8uLi4BVNLsfJ+RvJ0rd2D7v6DCxS0EeOBXmTe+epLJwetneL1aDZIbYzUVpdns7bP/S17r+ClU30gJ3zGfc0IQdPfQxXXGM1w7TgNGfgRoHNBsqKlkO4YNL3CB0Tizg2Vkf57D/V8zMdz9MfaPcZDzM0b3el0Nd5TrnuG5F8Fr33Gq6X7vdwjm9km3JSIXRbEAwFu6n2bHokWLzLqNqReBnRgd7P3MOIHn5EVe+fQqFgkpIziAzkLjLHN/afu49SzNlp1Tl0cYTV7ZQ717dSkw4t9sMuUZzkSo1A+ltpJKiYpiapPPrqXz3/8Vo6GYe7ggOHhQAWPraLp67+JWRv99n+Xxa8qY7Ez6jcU1lUWMyge+yASdoOZcTUdPRu111UDKJu5QBIGvz7tA2kBy5M5+3KEYIuMQF6vyPO4WJA65KINa8OOLSTkBvMENgw9tLY8REE8J3uBXgLVPyUVLcfdz7mlZPo/noUjoZR2jGufqcVJACQ8xgq8o5I4lTzf0uuQqC2jGfKnj5wVG+aXZvJ6DX2VyDuA1/3GMnE+4uI2UU5/Zxp/9zyB5lUyRFWUw6TvkdZmOqpNVSIsWLcKssf3ZmfDQt7xOAR25e2lOJw5QFZXwMFswlOVyIZaKo8J6MVAQtdyJGsDoXlfbAu0nM5pP2aLLpbTBjZmgZoKl+iSLplYsGn5xdA7LZ8jbvz7/YGQU3AWoKNJ34hH96XRdAjkEIvMIHVfMCEZ41WV8nqAmt27rRGqh1WA6hJ3vkS5I3WE87aYkW1fSb2CjWwjVHrH3suwc4CLkYWLuoF8cE7OekYyA7N1YJBQ9FNj6uvy88H5cGAqvUGGTnUz7yvOoarjvM3Ljag0d7pk1jM6TlgMQyW9e3Kr/3mF9uYtJP8AovzCduungboy2931OCWD0UDoiZdIVoD1BnYF9X9JZnFmr33UyeSXpGbdA5jicvNmNMecC8ONYmY/+42NGtQkzuNOQaBMJpdlcxA58w7/30/UYd/FnEjdtH+WOPZ4wTgbv/pDy0sbQeBUFQPYZUlFeUaR5JJQXGFfDiqKcp6mp4LlSwrc10OMxfq9qq7hQmZIONgdsbOvPN6nUAG4hmr52HFg0XpbH7vscSPy9mYdF3x2waEfu6tqEafZ3Aq1HATO3ofbqKdio1bwBA+NJHXR9lG1j0w8wUnQLBna9L0d47iHs/fHr40z89fobmwS5BTF5ee0kn3dhE98noj8jX18TxR5S0jJXkcBSa1gJWZTF7f6h/5H7vbSTMjSpPSxAB9BmFKM8G3vK6LyimVDc/wXti5/EBUutYXe+M2voFE79oh9l957NRGrpdVI5GYfp7KOHMwlbVcKCqYtb+PzQXoy02k0kR7/rAx7DM5Il9FJPkovbWAxUVQiM/RLY+hadfUA8qYWzaykdjBmhXyUpoeASP1vaPnL+EX2Aq0dNNAXbSBnf9WQmMlc/yUVSrGOU7teOVBAApO/n+21/m0lcgBLSnLP8rHs/VRxYQKOiyZJsyjo1DuTgD80H+s/hAgWQ6nL2M54L6uDJBTvhIb1k5417yD2MPxYKk/f6kf/p1zjU1QAnlprVkVuqT2oW1cqtorEceYvAwW/1p3e7hzI5VppNB3jlD0ArMhr73WBmY1hvUgYaZ2DdP6lgqa3UH54gYezXgLMPo8jqUn7JayrYD9w7mhSBjYbRtJMvuV87d+DSdkbxnaZzFJckh4y9l1tsqXHVofk87qj/sF3u/i/JZWYe5vZ414d8raCi0sPelbzlWgO6wMYemPwzOefsU3LCysYOuLSHSV47Z0bwdVVMeOVf5vEcfQBbO+qRF47TP+6gl6nZPvwt8MAqLmo5ZxidntVN54kayvd08KK8s8sMRst11ZQ4XtjE3UzbMZRCnlkLrHrM2P6RHzI6n7aMSeQLm7iASU5bqkUITCBvu8hEf2rlhBwAGPYWF9ab4eJOVpAe+JqLTPQIJiRjFa10U3cAP42TK1qdfICH1nIBljh+yQGakU9uEkQRWDyZ+Rkl2k+knvz/KVqkamXJkiWYPHmyuc2oH0UZ5JNdgxgVSgGXSwAd2JHv6bjK8khxDHmN3LVUwFOYzhtv+zvkt0+vlps9GaLgEpUo988HquyoW3YPYyfFYz/RCfd5hknR3PNUdoQGMRHZbiIXC6Wm/cxv/Bn9MblTQc2iqOJMRnxdZlAlUpjGaG/I65wWdHAeqZbJS4wLkQA6ldoqcpqH5utTIADQ+UEWXuVf5GIhaqlyqS7V6c4TIeZfhDD0TeYZaqvI57r4s2hoyOvAuY20M7QHzxlAO9U2XEiDunJaUcoWLnASRn7AiE5aXIqzGNEr2+z2eJw7jFYDec6XTKWNDh48R8qCsqyjfK5hv3GAapRus3iNOz+sG7X3K3difnF6VY16qCzQ7x1zfj0Tq6G9ZUce1oeOO20fAIEOXK3R5Saq2HDtwNeAtg6FbSbDPX5U/e9nITC61wWBuSJDRx4/hbs0Bw+TXPvthqX6JIt25MXFxTd/kjmhraMionMiaZOaCt4w93zIG7z301S23BjzpnPApdfoWG3sAJdg8pv+7egwSq/LU4YkRA4kpy5q6bTdQ5g8KrjE0mnftoB7MLDlDf3xaQN1JedJy6nbHfI6uXZlcUZNJZUcmUeoHCnOpKIhvC+jVmdfRsi7PyDnn5BImkiso9Pu/TRw+Hvyue0nMDJe/QQj2+6PsbfIlT18r4B4lpH7x+krfra+QdlmUTpweiWlZ8cXsYpPraFj6zuHxTxLH+D4NIA2uPjLOYMD35C2iehHXb2hOmbXh9Rte4Tz/F/5g6qh6GFyck5lw8XB3g0ovipHvV6t5Ik9SlzezbatqxV93FuP5CI54EU2itr7ObBd8XlHfkjnrjZx+5laHM+tZYtZrwj+rq2hIihpBf9fV8NzO3UZ1T0/yZJK95TNgPtK0m8WDJP3elhvYMpiYM+nXES7P8aE8qrHuTvq+aTxYAtz2GkBsGhHbvFwDeKW/7enZOdYUwGsf5Y3a+p22YkDjC41jqQCJI4zciD56tQdTJ5te4v/luWQW44cSCcpTUivLuPzT6/m+5/7nfzwgBeMZ2Du/Yycado+/jh5U9b4x3/49zajuZhsflWePt75YdITUj9wGzsuAFJ3xt0fMnLPT6UtZ3/jdPdD37LoaNsb8vvveI+vTdtLiWX/54Gk1XKzKSWu7GEiNv0AhGFvcRTchU3kQ3s/zSTi2d9Y1JS0kiqToz8yyj6+BMjYz6TpkH9RymmqD3pZDs9BQEc60e6PUcuu1jDaLS8A7vk3P1toD/3B0HkXucCl7tA/ZswIjodzDabU0M6FdEZZHq997gX9XuIAcxER/enkpUhagrLXjgSPCPYbkZCfyroCJS169QQXwhOLjF9/6DuLd+QmYefERTFyEKm474bK39OjP/D+mLb85tr9/wewaEc+bdo0c5vQMNQ2pB0MO8PVVPCGlqSJLv6U/eWnMqqVnHjbMUyeScnGK3vo+C7tZJn+kDfIv2efko/d+SH2IinO4iKyby0fN5XrqK2kI3YPZcQpRZ1D3+J4NkHFxUDZn8M1UL/XSG0VlTgxI7hgAHSwbsEcFNBvDotMBszVd+IS8lJZ3JOfyuq+oW8aDxYGGPlfPcHPcfh73qAXttApr1Fwy2n7eIxrJ3guD84D7vlAV1HpSQoDIO1kWFDS+h7udCIH8PeIftR+H/qO2/TgrvKCmXGYkkKps2VFAaPzyEFyVXDEQF43Bzde48WT9Rduz0iOijO8Nh0m8Rwmr+Ci0vlhOUEZ2IlqHakVrFrD/IayYEcUTV9vra4/jiFsjIcd3zHUVjIJX1VCtYyLv8mnNXiv29pzhyU5cQmZh7krvYOO3FJ9kkU78kOHDmHAgAHmNqNheIQbV/zZOgB2bqzUDOvDZGJ1CW9YiUiXqgWlPs+Ciu10q8tYtekRzqKd+EnAGQc61PYT2NBKauivdBoQyaUqByVIU8RDezBirKsGyq7zhq+r5ra8LJdyuYJLpCZMtSvNTyWVI8HZlxWlp5Yxqu/9FBcj91BZbSPBzkkeugFQfRI/laoXaaHTOHOAxSmOR6v1iYNNcHfuPH42MZEl/QDppOxkOme1hnSTEk6+wKiPKFkrTOfzvFpBb5aKrQMLdVoNIj3z41h5kdXWkvue/DOw5zMmklsNpsS0MA15+Xnwiu4mN1oruGxwPXTnzcFd/7r4tSMltGkuf888Sp39zK20zzMcGPk+KYTaSi4wOReohZfgGcGK1NOrFI9F8RzET2G0LiU7BRXQdYbxObwTqCjkrvCP/3ARdAvh+VSoaiTc9F63dTR+TKU27uh5m2GpPsmiHfn58+ct8qTdQFkeS98HvEAeV+LIB/8LgK5Uedf7skb81HJg0iJ5ZJmyWb+oZeK092wWovSezUh8yxluwx29gO3vksYAKJHzjaNETm3HCG7A80DaASYSO0whF6+sIvVrx2RecSaVLqv+KtvmEsARbaZ6e0cPoawOYGQYNYS9sIO7Uw2y/R3SBJ0fZoJRcmhOPtR3S1EuwM/uHkJ+E2Dvcu8Y2hQ/BQjrjRNF7uhsr9sl2JmQe9m5kG5oO456aFOaaAc3OvOQntxNXNKNEBtiYtcAMFKcspjXs/AyF7/gbqRcWg0m/SFFxc4+WPH7PMzqoEhMm6ratHeTj7vqSR63/UR9vhzQ6cZPy8UzgZ14vXNTuMPwaaM//ETjBAx7k5LE07/y+9F+Inclzn4cQpy8EtDW4bpfX/hK0sU7javHScVJKEonjTdpIZVLCtz0XvdpTVmrsgdOr6e4gN1BWKpPsmhHbvFI3cYKQtcgRrWCwCk6V08CLkVAxgHZUUo4+QsjRWXBioTS67ypKwoozQMYGUp6bycfRiaRA4Hgzpw3KaHLI0DWCX65pUh32+v6x89OItfoFkY7lLaVXJWjyv7PAQfm6eRvw+kwS64ClaNIV+z5LyPFzolcLJx9qb/OTqY2/noyHXZdLfudSxBUbMqVeZhJOdcg4OhPvKl7PcXPv+9LVLWZK7+m2yNMgEoRptqWKhZnX0BtT919fW2UD84Hgjoycms9EqitMY7clfCM4I8hGhP1+bRhtC7lH7rM4EKXto8U1CObgYp82rrzPeOCIUOZoHuo6Zayyr/3fII8v/K1KjV7fYfSea+aNw+zutx5dQcA7oQMcXk3v98GjvymcPBgPqTDX0iP+cUBQV24U7LCsh354MEWnKDRajmQAmA0uVs3litmOMvibWyZaTdE3gVy61FD6NiUEXBIN7md7fWz3PJf1PGxgsCWqVXF5ExXP6F/3CP/Y0HR6dV0wDEjTLeadfYjr2+YGAUo4Ws1iM7RI5K69cMLgGUPkdLp/ldGRp0TyeGXXuOC02Ey7S7KoBPPOsaqT20to8zh75I+cfblDiV+CrfZez6R33vV42xVkHMG3UKdqaaxtWdU/PAGytAEG1Z3inVU6FQV06nbu3EnFNCRyWSAycnUrfwBWFHqG0tqqRkGFBh9N+1cdMO4R5Bm2/4WcwcAF79py3hutVq5DYIE9zCqlv4MbqITN+s9ZKqKNaQHnbIBGmWnawAnX5kRluqTLNqR19bW3vxJ5oJKRUrg8m79x10C6NA0znRiKrV+dVrbMeQNcy8A477mVrPkGqVWMcM5cR3g1rjdeOD+b8lH11VRE355N9BmJDlzJUSRDaIyD3ORODiP/6ZskZ/j5MMWAYJA2aIhnx3WkwvNllf5Ge77Qh6EoK0B9n8ODHyF9qTtlxUcYX1YNbrm7/zcTr76jqrXU1Sq5KewZap3jOmhy1f2AGfWwGbXhyyA6jiF5y+km1zNd2UfsGCkvACmbmcfGm0NE6sJ0/m4NLjaxo5JywtbuAspvkpZZmPbFtcDk99Ne1dGwmd+Y7JUgraWVNTDG5iY6/ww5Cv8HwAAFjRJREFUqa1z68n1x4xoOPpubjvvFALiGXxI9J6zHzD8LZPRuEXf6wpYqp2WN6ZagZ07d978SeZEwgP6ig8HD5bTp+4iJ7v7P1RYxNxDRzRwLhN11aWsQsy5QCfU71kmAAvS9CN0lZqRpcaR0WZId0bBFYXGQ281TnTmYb3pJC/tZATadSaj6A6TmZRc/BdGwi5+jCBtHan26PccuwfueFduzXtyCbfuUuFFYAK5aVtHfRnelT/YEtU3lrz9cQMJ3P4vgMh+pJN2va+b+GJCZqdxludArn+OVI8hzq4z5vGTV3JhTN8nb+e9okgLJTzIIqzU7aSczq5lP/ASE3rtW0CD303DIiiAO6AaHZXl5MUFfdzXvCY+t681qlnvIUdPfudn7WQB06PbyP+bgMXf6zpYqp0WHZFbPAI7ATM2ofD8H3B382A23j2MlMSiidQtb5rLRlNdH2WUJmqZHPVvzyrM68ksVAGoABn0Mh2BX3tK1Ha+x2RbdSmP1+Zeasfv+YDRc8Fl4PJeFoxc3E7aQprevvczbm8DO5FT/30OH0/Zwoi0+CqPY+dMCqe6jPz4+U1M3kUNZq+W8fMZ4Z5YxkEVFYXG5yJtL/MEWUeM/6at05fLnVnNc7BwnLxbcfLmYiRpt6uKWeZvCFPVfGpbJnhDe8rqIQc3avnTD7CNghIFl/jjYmrwVTPA20SRSofJjEglZCdzV1NbxcRqQMdbmxnbUqBxZA95K24rLNqRx8U1bft7R+DXFskpBejdobf8mHdryukubqEDu7STPyM/pFb6j4+ZGINIRyZ18zu7lj8JDwGnVgAXN/N4zr7A5XOMlNuOYZvcdc8wOegeCtzzPlCaS8ebvp/RuiRRLMoAIOg7EYDyNvcQYM2TjLIHvsSe1XYu5LS1tXwPUaQKZ/dHdLIOHlwsLm7RP15oTxYztZ9IxYVyNJ1bMOkYCc6+TALO2MSSdxsNC16UnLl/ByZDDdFmFLfqytF0MSOo6Ikdq8/LeoSyrYAhBEHWm/9JNPjdDOgI/OVH7ipKr+t2Q0/Li9DVk8CCUbIkUW0LPPgbqa1mRou4h2C1s6mwaEfert2fTADdYRjZqXHU9e0+JCtIwvuwRWq6bu7G9dPk02vKSa1c2cfy7O5/ZUTt6AlE9ObrNc6U8eWn8qZfP0fWexemsbK0w2Q6ddcgRtkHv6VqJrQndwPLE/Vt9GnNqBBgdFqRz6RcVQkVKlLTr/xUKlmkSLmigM8LSpC7HgZ24vDhiztYfDLmc+4yMo8wuRU3DtjyLz7XyZuLisaJ7X7dw/j+nuVAZn/gwgaIob0gjPzAdKFHYALw4BqWqFeV0I4Ti9mQKnqIsYP2jqaCREo8Aqxu9WzVyKtrGg1+N23tuOCG9ABqK3idlQU959br6/3ratiCN7ir6bL922WnBcFqZ9Ng0Y586dKlFtnE3RAm7QzuAjy6nRN6NI50KHs+k/9+ejUweTGTdDVVVGf4tmF/c8lpDniJQwSkqeduIaykNCzaKb1OxywNnOg1m21a/ePo+CoKuUuQkpsu/ox4k1bIx6irZYe+vIuMli/tJG9enGU8yHnvp+Seez/DhSY/hQtBr79xkdnzCQtwOk7l+6YfYt8RQaADS/qVEWt2MvCLbjizoCJnP+R1bD96DoPqU3GodBJG72ie27wLtDsgAXA0HKsN2j7gJUbt+alMQAcmyOqWP4lGfTfro24k6kuJEl0LhGZ25C36HrJAWKqdFu3IWzy8WslFHgDQ7x9AqwEs/vCJZpR9eReLeS7tZDJTcuLOvpTKFSh02EXpLP4xnIto60DnXlXKaNzRg8Uk/vFsNbvmCTa7SniQ1ElNBbDhefn1oT0YPV/cygHCv84i5bPjXf6tzSguIkUKXbBHOBOPds5Uw5xYQv5f0sdL3RUHvcJFzdGbCVG1mrY4eQMrZtIRA3zt7o+AsD5ISc/GoJudWydv/oT3utkzmbOIGX7z590ptBnN3jRKdH/sjlcpWnH3wKIduaenp7lNaBQabaeTzqEonYpHJGkVtS2jWgmuwfrFNBIu7wZ6PgXs/S9/FwQqY2wdWf587nfg1BFG3GIdpX6CWr9/SvxkqmUu7yZ9UlUsTxBK208O3COcTaTKrgO/PEAJoY09OfewXizPL83W0TDXqS0/u87Y3rLrpGCiBgOxo+THizKpJDFEweUWcd2bZGNIN1Y3bn+X1EufZygVvQ1oCecSsNrZVFgHS5gb1WVMPK6cSfpB0pHbOlIFoixxBui0L2ymnK+mkotDQAdWVqbvpyOW4N+eFY3Xkqg2cfYjn+3sx78d+pYcu5KqGfMli43s3UiN7P9K/tuQ18jBF2cxWlf2Fmk3kdH2iSX69g56WScXVHHSjfJzL5kuN6CSMG1Z/T3ZbwfqainzrK869HaiogiA1mSBjBUGKMpgoVlZLqtoAztafI/124H6BktYtN5p4cKF5jahUWiSnRon8smTl7Dv97C3qYGWnG6nB3SORsU2qhonoN39lK3Z2FGf7OzLiK6qhLrx6GF8/rVT5II7TGLBTtw4+bi1VYyylZFgzAhWhZpSi4R0A+w9WIK+/wuWS3vp+lyE9wUC2gOBnUkXAUyIdpnBsnhbJ+ORahontpx1VujJEx4CAjrdmetelscCoR9Gs6o049YCi2ax0cHttjvxu+IeKr4KLJsBLJ3OiVTfjyBtZwZY6vm0aGqlvLz85k+yADTZTrUtEDWIPwAdNkQqVzwi2BSpspi8uraWFZTFmYzGx35NBUZ5PiOVpBW6svh3WHzj5Aec+JmOtSidjlY5nq3jNPZ+Kc9nP+0jC8hp+8aRXun0ABOpvm2ZxLx+hq/bOJfdGrOTWU26+VU2wJq2jEVLoghUlrBIx9mPdIy2Tr+kPLAjMHMbNeAaFyYw7V3vzHVPXimP6Evbx06Cj2zmTqUR+H/z3bxDaNDOayepwFJiwwtUghkWxt1mWOr5tGhH/v8WStmdfxx/RFHe/s/cSn7a0Zt9Q6rLqE0/9zv/nnGYDnfIa7pxZh04Laffc8ZUzfFFlC1K48VixwCRg+mcU7Ywak4/xAj+0i75dVXFHI7c42/c8saOAfrMBg5+Q3rFxl6mXgITKLP0ijauYnQPabiR1e1A6XW5N46Emgog63ijHbkVdxCmCsMqCuQqYCss25EnJiaa24RG4Y7YqeRwXfz1G/QXZchOXEJ1GaPwinxyiYKaXLWy74sEafiCoyd3Bevn0HGP+QJYOJ4OufU9pF6UMxQDOzNxGzsagEBpoH97OnIlf+4dzeRttYkbsryAXRnLcgGvSMC37e0/n4LKdIWoqSZn9cD63WxeNGinT2teL2UTuLj7m6X52a3CUs9nkzhyQRA+FgThbUEQXhQEYYkgCKbHf/xJ7Nq16+ZPsgCY3U4be2OtN0C98rVTbPfZ5RFGop6R+s9x8iGNMvAldk8sTGPb2rFfssCn199ZsOTZitz55MXsUjj1FzZEsrUnXSKVl0cP1y+2kUr993zMCUFnf2cuAKAT3/wKeerlicC8AcC59di3b99tOEnKz+xNbbkS9m5AkOk+IKZg9mveSNwVdvrFAdNXsgmXnQsbrw2aq19kdYdgqeezqRF5mSiKLwOAIAjPA5gL4O8Nv6TxSE1Nba5D3VaY3U73UI6I+/2f8mOhPakjD+oKLJlM9YpvHJUxJxbrZmQmsC3omifl3uRTfwGSX+BN4xHG3uN11UzK3fspX3tmDZ8bPYy6dc9w+X29o4GH1lA9U5xJ+ue3p3U69xJg6VTg4Y3sEng9mcOkJYhaYO1suPVUDMO4XWgzSvdZV7I/e+y9bPrVSJj9mjcSd4WdgsCxfA/+xl2ds6/pHdUdgKWezyY5csmJ66ACYGLvbMVthyBwJJx3NEv/XQOZ1HQLZtXoQ2uB3PPIrQC8I/qzIKUin5HxtrfIX2uc2KnOI4L8u60D8MdHcuFRXQ2TTpITB9jUK7wvO/gp4RbMsXS5F4D/DdOXN4oidwmh3akcMUR5HpzUJqYUNTfsXY01/VZYNhzc9CclWXEDN3XkgiBsBGCq1vhVURTX6J7jDmAYgPENHGcWgFkA4O/vj3nzWIDSrVs3+Pj4YN06FpOEhoZi2LBhmD9/Pqqrq7FgwQIkJiZi1apVuH79OgBg/PjxSElJwYkTHOLbq1cvuLi4YOPGjQCAyMhI9OvXDwsWLAAAODo6Yvr06Vi+fDny81l0M2nSJCQlJSE5mf1G+vfvDxsbG2zdykEEMTEx6Nq1KxYtYktWV1dXTJ48GUuWLEFxMftkTJs2DYcOHUJ1dTXmzZuHwYMHo7a29kary7i4OLRr1w5Lly4FwGKCCRMmYOHChTey34mJidi1a9eNlX748OEoKSnB3r0crRYfH4+oqCisWMFyel9fX4wdOxYLFixAdTUpipkzZ2LT9j1IS0uDWq3BiBFdkHMpBwcPkjfv3LkzgoOHYtmyZdAkrUZgYCBGjx6N+fPnI9D3XsQMn4ComLbYduwirm48iHv7vwHXogz96lGvVqaHJp/9Devyw5GTl290nabcPwrOLoEQlA2zABRUC1g2bx7G9YyCj2G/9qAuuF6pwnbd96M5r9P58+cBoFmuU3x8PJKSkm79Om3ahLS0NADAqFGjkJOTg4MHDyquUzBWr14NAHrXSavVQqVSYebMmVi7di2ystgUbcyYMcjIyMCRI+w6aXg/eXh4QKvVYv78+QAAjUZjkfdTdXU1ysrKmv06/en7qZ7rVF1djSNHjjT7dVL6vYauU31ockGQIAhuAL4CMFcURROliMZobEFQUlKSxTapUeKus7O2msVFP9wrP2bvTp23VFEqof+LwMAX6j9Wylbg54mys/aOoUTRI5zFOCmbSb2UZrPfzL0fIynHcpsTSbjrrrmZYbWzcbgtBUGCIHgD+ALAHFEULwmCUP+S8ScgraKWjrvOThsNO/Hd829Z9y3WkYYI6SE/z68dKZSGENEfeGQLpw1N/AGYtpxOHGCDqNb3cPDAEweA6SsAv3Yt4ny2BBsBq53NDUu1s6nJzk26YywSKI8rAbCiwVdY0TJg60B1QHhf3cDmEA4mnrwIyD3PxKR3jH5lpimodS1vgxLqf45rAH+ssMKKP4WmJjsbuDubjvj4+Nt5+GbDXWunjcZ4tqXUdfA2oiWcz5ZgI2C1s7lhqXZadEFQVFSUuU1oFO46O2urWdmZupP66oh+HI5QlgNUFnB6UVEmBx2H96UCxBx2NhUl2UDWURZUeUWxB42DiZ7mJnDXXXMzw2pn02DRTbOkzLKl466z8/JuTqrf+R6w8UWOJTu5lB0UfxrHronnfweWTNWv9LzTdjYFlUXAppeBxZPZc+WnscCBb/SrBxvAXXfNzQyrnU2DRTtyK8yA6nJg5/v60sPKQk6wqak0nk6089+s0GwIWi2Qcw64sIX9TAyPYQ7knAVO/aL/2K732cDLCitaGCyaWvH1vUkizUJwV9lZV0P6xBA1FabbAKAR8tWUzRxOUVvF4qXBrwHdZtU7bu2OnE9TjZi0tVzIGoG76ppbAKx2Ng3WwRJWGOPoj2yVK0EQ2AJAW8suisqIetw8VpXWh8J0YF5/Kl+UeHR7w0qW242CK8C3A/Xt8u8APLiazcOssMIC0SIHS0iVZJaOu87O1qOA+z5nf/Ow3sD479kTZf+XLOPvOJ2dECf/TB14Q6jIN3biAIdEN9XOpsAjDJi2AogcyEZMcfcD93/baCd+111zM8NqZ9Ng0dSKVDJr6bjr7HTyAhIeANqOZStcQcXGWIXpbJ/bZQZ7szQGzv7Gg5sFFRt9NdXOpiKoE3XxlUWAo9ctddO76665mWG1s2mwaEduhZlh7yL/P7ATf24VLn7A+PnAsocYhWucgNGfsJjIEqBxavyiZIUVFgqL5silBjSWDqudjUBxFn8cPFkh2sCw45ZwPluCjYDVzuaGue1skRz5pk2bzG1Co2C1sxFwDeQsUK/Im06sbwnnsyXYCFjtbG5Yqp0W7cilNpKWDqudzYuWYGdLsBGw2tncsFQ7LdqRW2GFFVZYcXNYNEeemZmJoKA7P2D1VmG1s3nREuxsCTYCVjubG+a2s0Vy5Dk5OeY2oVGw2tm8aAl2tgQbAaudzQ1LtdOiHbk0AsvSYbWzedES7GwJNgJWO5sblmqnRTtyK6ywwgorbg6zcOSCIOQAuNKIp3oDyL3N5jQHrHY2L1qCnS3BRsBqZ3PD3HaGiaLoY/igWRx5YyEIwmFTxL6lwWpn86Il2NkSbASsdjY3LNVOK7VihRVWWNHCYXXkVlhhhRUtHJbuyOeZ24BGwmpn86Il2NkSbASsdjY3LNJOi+bIrbDCCiusuDksPSK3wgorrLDiJrA6ciussMKKFg6LHywhCMLHAMoBlAKIBzBbFMX654SZCYIgqAA8CuBNAINEUUwys0k3IAjCEAD3A7gOQBRF8XUzm2QSgiD4A3gLQLwoil3NbY8pCILQCrTxKIBgAHmiKL5hXquMofs+/gbgAAANgFYAZoiiWNHgC80AQRAcQDs3iaL4rLntMQVBEPYDqNT9WieK4mBz2mMIi3fkAMpEUXwZAARBeB7AXAB/b/glZkE8+GVs3Bj2OwRBEBwBfA0gThTFKkEQVgiCMFgUxa3mts0E+gBYDaCjuQ1pAJ4AloiiuBoABEE4LQjCOlEUj5jZLlPYJ4riWwAgCMJqcDFfZF6TTOItAMfMbcRNsEEUxdfMbUR9sHhHLjlxHVRgZG5xEEXxGAAINxmaYAb0BHBFFMUq3e97AIwCYHGOXBTF5YIgDDC3HQ1BFMVDBg+pAJSZw5aGIIqiFnSQEATBBtw9nDOrUSYgCMID4HeyAwBnM5vTENrrAkkHAIdEUVxnboOUsAhHLgjCRgB+Jv70qiiKa3TPcQcwDMD4O2mbEo2x0wLhC6BE8Xux7jErmghBEMYB2CiK4llz21IfBEEYDuAZAGtFUbx57+g7CEEQ2gKIFUXxJUEQOpjbnpvg36IoHhQEQQ1glyAIJaIo7jK3URIswpGLoji8ob8LguAG4EuQ48u/M1YZ42Z2WiiuA1BMUYar7jErmgBBEAYCGAhgtrltaQiiKG4EsFEQhB8FQXhCFMUvzW2TAuMAVAqC8AJIq2kEQZgtiuInZrbLCKIoHtT9WycIwm7w2lsdeWMhCII3gE8AzBFFMVMQhPGiKK4wt10tCPsAhAmCYKejV3qDi6IVfxKCIIwC0BfA0wACBEEIE0Vxn5nN0oMu2o1QUACXAESa0SQjiKL4tvR/QRDsAThbohMXBKENgN6iKH6neygawEozmmQEi3fkADaBdi7S8c8lACzOkQuC4AHgSQBuAGYJgvCzKIr7zWwWRFEsFwThcQCf6rpOnrTQRCcEQegP4AHQOb4M4CNLU1kIgtAZwFIAhwFsB+AE4AtwwbQkVAF4RBCETgBsAcQCeMq8JpmGIAjjAfQDI/IpoiguNrdNBigGMFoQhEBwR5sOwKJstFZ2WmGFFVa0cFgLgqywwgorWjisjtwKK6ywooXD6sitsMIKK1o4rI7cCiussKKFw+rIrbDCCitaOKyO3AorrLCihcPqyK2wwgorWjj+D8+UrTfBmEMDAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_pca(X_dev, y_dev, path=\"analysis/analysis-results\",title = \"original\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAEPCAYAAABfmE8WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOydeXxU1fn/32f2JZnsCwmEsAgIKKgIKgUEFeryVSogKn6VutVfxa1fLbVaqxVrrVVb676iooCCoFKtCgVUlF3ZZA9LFrJPMllmu3PP7487M8lkEkggIYmd9+s1r2TunHvuZ+7MPPfc5zzneYSUkhgxYsSI0X3RdbaAGDFixIhxYsQMeYwYMWJ0c2KGPEaMGDG6OTFDHiNGjBjdnJghjxEjRoxuTsyQx4gRI0Y3J2bIY3Q5hBBzhRCyhcfDwTYHm2z3CSHOO0a/q1roc26jNk37bfw4vxXazz/K/qGHIoQoEUJ8KoSYLoQQJ3jKYvyXI2Jx5DG6GkKIbCAJuAKYE9w8CSgCSqWUpUKIAYCpSZtC4AwpZVkL/fYB7MDZwBvAjcAGwCmlLAy2CfU7J9h3UfDYAAeklHXH0G4H+gSfbgv+fRF4IdQE6AWMB+4CjMC/gClSSu/R+o4RoyUMnS0gRoymBI1qoRBiRKPNe6SUBxu12QPQpE028K4Q4udSSrWZfg8E90kNbjogpdzepE2o36rgJn/TNsfQXgdsD/YR2lzapI9twKdCiJVoRvxS4Ang7tYeJ0aMxsRcKzF+KnwY/HsR8MfOFNJapJSfAt8En94qhLB2pp4Y3ZeYIY/xU2E28HXw/weFEJOO1rgLsTX41woM6kwhMbovMUMe46eCAlwFHEH7Xr8rhOjVuZJaha/R/zFXZ4zjImbIY/xkkFIWA9MAP5ACvC+EMHauqmMSGoUHgH2dKSRG9yVmyGP8pJBSrgHuDT49B/hbJ8o5KkKI09F8+gCLpJTOztQTo/sSM+QxfnJIKZ8F3gs+vVMIMa0z9TRFCNFDCHEt8G9AD+wAZnWuqhjdmZghj/FT5RYa4rhfD8aHdxYPBRcBKUIIBS02/V00F9CjwEgpZXkn6ovRzYlNrsT4SSKlrBdCTEFb8JMALBZCjJJS1neCnJdoWBAE2sRsRUsLl2LEaCuxEXmMnyxSyr3A9YAEhqKtsOwMSqWU2xs9dsWMeIz2JGbIY/ykkVJ+DDwefHq9EOKWztQTI0ZHEDPkMf4b+APwZfD/Z4Ez2rNzIcRMIcTM9uwzRoy2EDPkMX7yBPOuXAMcBiw0jNDbi5nBR4wYnULMkMf4r0BKWQFMAbyAuZPlxIjRrsSiVmJ0ORqlsc1utHmAECKO6DS22c293ly/UsqNQog7gFeOcuzRaHlPegQ3WYQQFx5DchLgDO7fOI1tiHQhxNDg/wVSyipixGhHYvnIY3Q5goUebmjh5UeklA8LIQ4CvVt6/Rj9vwbcBIyXUq5q8lpL/R6L1VLK84PFJ1Yepd0vpZRzj6P/GDFaJGbIY8SIEaObE/ORx4gRI0Y3J2bIY8SIEaObEzPkMWLEiNHNiRnyGDFixOjmdEr4YWpqqszNzY3YVl9fj81m6ww5baY7aYWY3o6kO2mF7qW3O2mFk6N306ZN5VLKtKgXpJQn9EAb1f8LeAiYA8wHrEfb56yzzpJNefnll6O2dVW6k1YpY3o7ku6kVcrupbc7aZXy5OgFNspmbGp7uVa+k1L+SUr5IGADrmxrB5deemk7Sel4upNWiOntSLqTVuheeruTVuhcvSfsWpFaHos5AEIIA9AT2N20nRDiVuBWgMzMTF55RVtcN3LkSNLS0vjggw+w2Wzk5OQwceJEXnvtNQBMJhMzZ85k6dKllJZqC/amTJnCvn372LJlCwDnnXce8fHxfP755wD07duXsWPHMnfuXABsNhvXXXcdixYtorKyEoDp06ezfft2duzYAcC4ceMwGAysWLECgAEDBnD22Wfz7rvvAuBwOLj66qtZsGABxcXF2Gw2ZsyYwYYNG9izZw8AF1xwAYqisHr1agCGDBnC0KFDWbhwIQDJyclMnTqVefPmUV+vpcWeOXMmX331FXl5eQBMmjSJmpoavv32WwCGDRtG//79Wbx4MQDp6elMnjyZuXPn4vNpdXtvvvlmvvjiCw4fPgxoX6iysjLWr18f3sdgMPDRRx8BkJWVxWWXXcZrr72GqqrodDpuvvlmli1bRlFREQBXXHEFBQUFbNq0KeJz+te//gXQoZ/TZ599hqqqJ/w5uVwugA79nJYvX47NZmuXz+mss86iZ8+eHfo5+Xw+Zs2a1aV+Ty19Trm5udTW1na531NLn9OiRYuwWq0d+ntqkeaG6cfzACahla565FhtY66Vk0tMb8fRnbRK2b30dietUv40XCtIKT+XUv4c6COE+HV79RsjRowYMY7OCRtyIcRgIURj59ABoG9b+znrrLNOVMpJoztphZjejqQ7aYXupbc7aYXO1dse4Yde4CYhxBmAETgVuLOtnfTs2bMdpJwcupNW+O/VW+v141ckSXZTu/TXHP+t5/Zk0J20QufqPeERuZRyv5TySinlHCnlH6WUV0kpi9vaT2jioDvQnbTCf59en6KyencpM15dx+XPf8PcNQcoq/G0al9/QGV/aS27il3U+ZQO13qy6U56u5NW6Fy9sXzkMX5ybC2oYubcDYQSez78yY9I4Jejm6YJj6Si1svr3xzgla/yUFTJpMEZ/P7SU+mdYu940TFinACdksZ2xIgRcuPGjRHbli1bxmWXXXbStRwP3UkrnBy91dXVlJeXh0O4TgS3243Vaj3u/Ws8fqrdkaNpg06QHm9GpxMtH9cXoKIuUr/DYsBhNXaY1pNNd9LbnbRC2/Xq9Xri4+NJTk7GbG5d0SohxCYp5Yio7V3FkMfovng8Hg4fPkzPnj2xWq0I0bKxPBlU1HoprHJHbLOZDPRJtaHXtexNLKpyU17rjdhmNujpn24/6n4xYrQVKSV+vx+Xy4XT6SQnJ6dVxrwlQ95lvp2hQPjuQHfSCh2vt6ysjLS0NGw2W7sY8bKysla3VaVEbTIYsZsNGJoY3gyH+ZjG2GLUR22zmfTojvKe2qK1K9Cd9HYnrdA2vUIITCYTqampJCUlhRdWHS9dxpCHVvJ1B7qTVuh4vR6Ph7i4uA49RlNUVVLj8XOwvI6D5XW43H4CqmbQLUY9fdPsZCdZ6ZFgoV9aHHbzsaeD4sx6bKaGdgadjtQ4U6ffYbQHAVWlzqugt9jx+gOdLSdGIxwOBzU1NSfUR5eZ7NR1o1vX7qQVOl6voigYDCf3q1TvUzhQXhd+XutVyE2xh/3ZFqO+2RH20TAZ9PROseH1B1AlWIw6TIa29dEVUQIqxdUeKus1/7/e5Sc31d6qi1uMjsdoNBIInNjFNeYjj3HC7Ny5k1NPPfWkHjO/sh5nfeTEZLzFSG5K+7h3fkrUePwRFz3QXEa5qfYoF1SMzqG1v6Eu7yNftmxZZ0toNd1JK3Q/vVVVVcds01zwiV5w0o14a7R2Nv5A9GDN7Qugql278Hp3OLeN6Uy9XcaQh7KEdQe6k1bofnr9fv8x2yTaIn3XAkiOa10IV3vSGq2djckQ/TOPsxi7fCROdzi3jelMvV37k4wRowVsJj39Uu2kxZtJizPTNy0Ou+nk+bMDgQCZmZkMGDAAIQRWq5XMzMxmHyaTKZwC9qTpUyU+RUVVJVajnh4J1nD0jcWoJzPBgr7Jbc3bb78d1tsRcx7H2/8NN9xATk5Ot4hiWbx4cYeewxZpLiViRz+aS2NbXFx8wikeTxbdSauUHa/3xx9/bNf+fD5fu/bXkXz55ZcSkDfccEOLbcaNGyfffPPNNvWrqqr0KwEZUNU2a6rz+uX+0hq5taBK5pXVynqvIlVVlW6fImvdPulXAkfdf/To0VKv17f5uK2ltf2HvgeXXnqpdDgcsqCgoMM0tQeNv7dtPYet/Q3R0WlsT5SCgoLOltBqupNW6H5622N1aJuQEhQvBNp+3I64nfb6Axyp9rCvrJZCpxu3r/URDT5F5WB5PbVeBSm1EM1DlXUoqtSieAI+DPou87M/KqHvwSeffEJpaSnZ2dmdrOjonPTvbSO6zCcaqpzRHehOWqH76Q1VezkpKD5wFUHpTijdBXXloLbecHq93mO2WbFiBTfccEPr5KgqBU5thalPUXHW+zhYUYdPabIWQFXBVwf1leBxQUC7oPiUAEqTdQMmvQ5/QNt2Us/tCRLSKoRo9RL2zqQzz22XMeQxYnQKbifUlQISZACq8zUD2Q7MnDmThx9+GL1eH56YVVWVf/zjHwwbNozk5GSSkpIYO3YsS5YsAbQRdZ1P4a4br+X8Yf0Z1iuJB34ziw+XLGHEiBE4HA6GDh3KkvfnQfkeVn26mPN+NoaEpBROP/00/v3pv8LHN+h0ZCVa2bRhHf97wy/p1/8Uhg4dSlJSEhMmTODLL79s0/s5lva2kp+fz7Rp08jJySE+Pp4JEybQOCy5vr6ezMxMEhISEEIwb968iP3nz5/PFVdcQW5uLmlpaWRkZHD11Vezd+/eqGNVVFRw55130rdvX3r06EGfPn24/PLLeeutt6La7ty5k6uvvpr09HSSk5PJzc1l1qxZzfroCwsLufbaa0lKSuK0005j7NixrF279rjOxwnRnL+lox/N+ci///77VvuTOpvupFXKjtfb3j7yurq6iOdLNhfI8x5fIXNnL5PnPb5CLtncTr7SgF/K4h1SFm6OfFS3vv/PPvusRR/5DTfcIP/4xz9GbJs+fbq02+1yyZIlMhAISLfbLf/whz9IQD7zzDPS7VPk1nyn3JLvlJ98tUkCst+AQfKee++THo9H1tXVyZ9Pmij1er2c/8Lj8sG7bpaevLWybt8aOfHCCdJgMMi1W3bJLflOWVztltsKquQlv5gm+w4YJD/+aqPMr6yTzqoq+atf/UoKIeSnn34apbsl/+6xtLeW0aNHS51OJy+44AK5bt06KaWUhYWFctSoUdJms8nNmzdLKRu+B6+++qoE5DvvvBPRT3Z2tpw2bZqsrKyUUkqZl5cnzzvvPJmSkhLlTx83bpwcPny4LCwslFJKWV1dLa+//vqo97lu3ToZFxcnL7/8cllaWiqllPKHH36Q/fv3l3379pUVFRXhtk6nU/bp00f26NFDbt68WdbV1cnDhw/LiRMnypycnJPqI+8yhryrT2Q0pjtplbLj9ba3Ifd6veH/l2wukIMe/Ez2nr0s/Bj04GftY8zVgJTl+6MNeW1Zq7v44osvJCAtFovMyMiIeFgslghDPn/+fAnIhx56KKqfM888U1qtVllWXi4LKusiDHlaeoYsqqwJT3z+e9lHEpA9e2RI/6H1Yd2fLf1AAvKvf31Sutw+WV7jkVvynfLmWb+Rz721UG7Jd8qt+VXS41ek3++XaWlpcvTo0VFamjPkrdHe2MgdjdGjRzdrmLdt2yYBef7550spG74HLRnyiRMnypKSkohtP/zwgwTkAw88EN5WUVEhAXnfffdFtC0tLZVDhgwJP1dVVZ566qnS4XBIp9MZ0fbjjz+WgPztb38b3jZ79mwJyDfeeCNC765du6QQ4r9zsjNUTbo70J20QvfTW11dHf7/yc93426SG8TtD/Dk57tP/EBCB/EZ2t8QehOYW583pq5Oc8NMnz6d4uLiiMf06dMj2oZcAxdffHFUP6NGjcLtdrPyP//BYtST4bCQEEw3MPzMs3B5G5KD9eqt5VU/Y+jARiFuIry9oCCfeIsRY3BS847Zf2DMhIkA6HQgEBgMBvr378/WrVtb9T5bo33FihWt6ivEhRdeGPF86NChZGdns3r1apxOZ8T3oDk+//xz0tPTI7aFVkc2fl92u534+Hhef/113n33XTwerchIWloa27dvD7fbtGkTO3fu5NxzzyUxMTHqPULk4rqQS2niRO3chvQOHDiQXr16HePdty+xZAsxujRFTdLRHm27ElDxKiqqlJgN+mYXwkRhskPqAPB7QAgwWsHQMRNru3drF5/LL788Kv+Nz+fDbrdzpKgIIwpFLj+1wQpFNkcSiVZ9eDm9yaLlvE5NS9N21psgoRfmuhKg4eJiMeqwGPUUFZfw3hsv8/V/vqCyrFQbwQmorKxsdUK1Vmk/cgSAWbNmsWjRoog233//PT169Ag/1+l0ZGRkRB0nOzubwsJC8vLyyMnJOaqmHTt28PTTT7N27VrKy8sjFog1nng0m80sXLiQG2+8keuuu47bbruNiy66iGuuuYbJkydjNBoj3uNXX31FZmZm1PHsdjsVFRXh53l5eQAR7ytEVlYWhYWFR9XfnnQZQ36sD60r0Z20QvfTazI11NjMSrRG5RYPbW+MT1EprHJT49GiN4x6HbkpdqytWSRktGqPVuBVAigBiUEnMBl0R1300dIioK+++opBgwY1v5NUUaqLyYiLJz9oY61GPUmG6DBHnSUB0gdrdxR6I4jSiNdNBj2pFsnPr5xEWVkZr86dx/BzxuBTBQadjpunXcz6deta9b5bpT3Ic889x3PPPdemfkOERsuhNK8t8cMPP3DuuecyYMAAFixYwNChQxFCoChK2DA35uKLL+bw4cN8+umnfPjhh3z44YcsWbKEYcOG8Z///Ifk5ORw28svv5wFCxa0WrPf78dsNh9Vb0fTZVwroduT7kB30grdT6/D4Qj/f9+kgVibZDG0GvXcN2lgxLZ6nxI24qDV3iyv9UblKj8Rajx+9pXWsr+slr2ltbjcfmz21peBCxnA/Pz8qNcURWH58uVUVDoxmMykefPJtWtxyTbqMembySEjhHb3oG+5gtGqFcs5kJfH5VOvYfDIsfhUrR9FVWnLqWmV9kaj1WOhqiqlpaVR248cOYIQgj59+kR8D5ryxhtv4PF4eOihhzjttNOOmWMnEAhgNBq54ooreOuttzhy5AjXX389W7Zs4fnnnz/mewTYs2cPGzZsCD/v27cv0JACo7He0N3JyaLLGPLuVKyhO2mF7qe3vLw8/P/kM7J5/MrTyE60IoDsRCuPXzGIyaeYoGI/uKvAV4eBAPomP+Y6n9KQGEoNgK8e/G6Qbc/P7lMCHK6sD+c8V6Uk3+mm3nPsOPIQ1113HQAffPBB1GsfffRRg9vC7EDYUzF5tPMgDBbNBXQctDRKlFJS0ILBao5Wa28Dy5cvj3i+du1aysrKGDduHElJSRHfg6aE3ldTA37w4MGotvv27SM5OVmL7ggSFxfH3XffDYDT6QTgzDPPZPDgwaxbt47Dhw9H9KGqKtOmTWPx4sXhbVdeeSXQMAcV0rt3796o/TuaLmPIY8RoiclnZLPmdxM48JdLWfPbsUzuE4CaIjDHg6sQyvcQ59pP3wQZkZbVYTFq+UQULzgPQPluKNsFriPhBTSgraSsdvup8fjDC2dQVc1vrmiG2h+QYSMeQqtO1Pr3MX36dKZPn87cuXN5++23wzmoV61axa9//WseffRRkpKStBF2XDokayM+DGbQHV8emTFjxtCrVy8+WbyAjd+t0d6vx8Ozf3mEgoLWG/Lp06cz7aqrmDt3Lq+8/gYen7957a1ECMHf//73cNz4kSNHuOeee7DZbDz99NPH3P+qq67CYDDw6KOPhn3VJSUl/PrXv262vcvlYvbs2eH5g5qaGp577jkMBgPXXHNNWNObb76JzWZj5syZ4ZF5VVUVt912G3V1dfzmN78J9zl79mz69OnD448/zg8//ABoq6hvvfXWCFfNSaG5UJaOfjQXftjWXBSdSXfSKmXH623v8MOyssjwv4CqylqPX5ZUu2WZyy3d1eVSVhVIeWRrZNhg0Q+yosolt+Q75b6SGunxKVKqqpTVhdEhhvVaeFmd1y+3F1TJLcHY7b0lNVoYWcWBYJ9bpKwtlT6fT25r1G7zwXKZkpYuE5OSIsIPzznnnKO+t0AgIJ977jk5fPhwmZSUJHv27ClHjx4t33///Yh2M2fOlKmpqRKQVqtVZmRkyNLSUnnnnXe2evvevXullFLu3r1b/uLKK2V6eoZMSEyS/QcMlH998m9y5MiREpAZGRnyySeflG+99ZbMyMiQRqMxvD0UblfvVeT2Aqe8/9G/yoFDTpMJiS1rb4nG/ev1evnjjz/KSy65RGZnZ0u73S7Hjx8v169fH25/6NAhmZGRIR0OhwRkQkKC7N27d/j1L774Qv7sZz+TCQkJMiMjQ44aNUouXbpUAtJkMsmMjAy5Zs0a6fF45EsvvRSO787MzJQ9e/aUkydPlmvXro3SuWvXLnnttdfKjIwMmZmZKfv16ydvv/32cAx6YwoKCuTVV18tExISZHp6uhw5cqT87LPPwiGWGRkZ8s477zzmuTnR8MNYYYkYEaiq5Ei1G50Q9Ehs3QRgRxeWaFwYQa8T9IzX4dB5EaofVEVbVh8kkNQPr96mTUTqdKgBBV35Hgg0cYHY01DjszlUWR/hWwfISTSR6NoDNPw2ZEp/XKqZ/Eo3qpQIBNlJVhJtxqPW9GwLUkrqfQEq63yoUpJsN2E3GdA1l3wdtLsGqYK+Y2MWpJQUVrmprIvMJdIzyUqyvesvne8O/GQKSyxdurSzJbSa7qQVWq+31OXh7yv2csHTq5n096+Y++0BqutPfo7lkM8StHSsZTUNRjjXIUiozUNU50NNMXhrwd4QS6zXG7CZDAgE1W4fBVVeAsZm/MtGKwEp8TRTv9KnqJGx5YDw1eOwGOmfHkefVDunZMSRaDNS3Y7FBOp9AfLK6nDW+6h2axevWq8S3VBK7X1X54O7AurKtOetyBHT+Ny2FlVK6r3Rfbt9Aeq82iRzVC6YduB4tHYmnam3yxjy5mawuyrdSSs0xBUfi+U7S3h2xV48fhWXR+Hhj39k/YHWRyJIKQmoKk3v8lRVM5heJRD5mqoGsw5GGitFUbSRprcWXV0JPc315Dh0xJv1WD1lkZOVSjD+GwHxWWCwaO/Zq3Coop4qt596YxJS1yiywxQHpjgMOhFedNMYqwEt70pj9FohC4tRT7zFiMWoRydE1Hs9EVweP5LI/pqNvFE8UHVI8527iqC6ACr2agb9GHHhitLMheEY6HU6EmzR58lo0JFXVseB8jryymubvSieCMejtTPpTL1dJo48RvtTVuNh0yEne62DWb27lHpfAINeMDAjnpyUyFGqTwmwcEP05NeXO0u4aEj04oimeP2aS8DlUYgz60mJM2Mx6vEqAYqrPVS7/STbTNjMelQJCcYAhrpihKca9GZI6KlNXobcFN4aqMxDACbApDNgcvRBV+OJPriU2qIeg1lbuihVjKqHgQ4FqdNTE9DhsfQm3hDQUrkaLKA3IoCUOBP+gEq1249OCDIcFmym4Ig8dMEw2sBkQ0ptdaUOifDVQX0FiQjwmrSoEiGCKXE92mSqzqAdq70r8fjdYE2E2pLI7TVHwOwAk619jwck2oy4fQFcHj8CQbLdhMcXCF94QpkaeyS0zh0Xo33pMoZ8ypQpnS2h1XQHrV5/gOf+s4+3vjvEPRcN4HcfbuNItWYE0x1m3rlxJAMzG+JeDTodAzLj2VIQuSy6b+qxl6urajAUL7gS0asEqPUq9E2Lo9rtp9rtJ8FqRKcT+AMSEdSnUyV60PzXlXmQNhCMVpITE8CZ1+QgCkbpR1oSEXVN7ojMcZrxChlRvwervwbclSBVzKY4fLYeVPjMZMVFGhqzQU+vJBsZDhUhtJSvQgjtwqB4NYNutOBV9VS6PLjcCnFmHSkGBYunGh0SPE5IOUVbVOSrBX+9tq/bCY4ssKW1ypg7LEbKa3wRo/LUOHO0D17oANF8GKV69FFhWyJLGmM26OmVbMOnaOfJWe+joi7S7VbnDWgXunaaMzherZ1FZ+rtMq6Vffv2dbaEVtOltNYUa7m06yJdIHnldby99hDp8WbKarxhIw5Q6vKy5PvI5cM6nWDmubmcnp2AMbj4JMNhZsKpkbksmkNRZdiIA8SZdOTYFQxVB0nxl9AvQRBv0cYMJS4PxS4PeVUByvXpqIbQ6DFohAFF8Ue7NgCDTqvNiSWYB0PotDA9odNcCvUVyMqD+D21qGoA4nsAAuGrxRyoJd3axEURzOmt81RiCdRiFipCVTQjrDdpo16LAwUDBU43ZTVevEqAijo/B2sN+G2Nzo3i0VwclXnaZ6J4ID5Tc30ozdxFNIPNpKdvmp1km4kEq5E+qXbizM2MtUKrUHVN3R1C030UQisnjwe9TmA16bEY9VgM0eGQCdb2m/iFE9PaGXSm3i4zIt+yZUs4MU1Xp0tolRLyVsFHt2ux1GmnwuTnIfssQMs7IiVkOCwUOqMT3v9wuEoLWwr+8PLKall7oJJTMuKZcU4OPRwWeqfY6Z167IUojX+7OiHIsipYag4BoAfsHici+VQKnZFRD6V1ARwJqVg95ZrhFppxqK33YLanae8r4jg6LXe40QZxGdo5cDvBaAfFjTcgqTBkU+2RWA2CdAE2WzLUV4DOgF42Ga16qjRfM2hGMT5Tc1eY4zWDqHjBZEca4qjzRe7rC6h4dVY0Uxp0qbgrGxr43WCwam4jtXUTxkII7GYD9ibGu/HnBGguJFuK5s6pLtAqG+kMkJhzzDwxbrebuLjWJwWLQA1oxxJ64iwGku2mcCRLgtVIgrV9zckJae0EOlNvlzHkMdpIxV5YcI1mMADKdsL7N8DNKyA+g5wUOyNzkyl2ublieBYrd0cmxZ80NDNsHIqq3Nz81kbygiF+izcXcPcFpzD6lLRWSdHrBAnBH3W8WYfZU9ykhUQNKFETefFmHSZdQFvsojOHI0WklGBJ0BrVlWtGKj5TM66WBPBUa+4L0Iy/wUIgoFDktVATjK7wB6DOL+iflILZYAZPNWFTaIrTXBDVjUrg2VO056HXfLXaJKy7EoMpnhRrBhXuyLsEIdC0CV10eCNoBSqM1mOOklvC4w/grPdR7w2QaDMSbzE2JAIzmLWH0dbIH9+BuT787oaiG0KPMaEnWQkJpMaZkRJMBl1UMecYJ48TNuRCiH7AHGAz0BOokFL+qa39nHfeeScq5aTRJbQ6DzUY8RDV+Zoxis8gwWrkL1NPY31eJXtKaph+di+WbC5EIrl8WJaWv6S+AqGgdMQAACAASURBVPLXs8uVFTbiIV5cvZ/JZ2aTm3LsEXlokjDObNCy6TWTsNAkVAx6HUpw5aReCLIsPvTVjSZY3VWQegrJFhUq9mkrHBN7BUe2wa9qfA9tAtFdpRnJuAwwWvCpXmq8kQcOqBKvqsPsKmrY6HVBSn/t/8buGykBCbakhtqdZjMIgagtIT0pg4pg9wJBbqIeG/VIaxLCFN+sKwijTbtASKlN3uqMmvFthfvBpwQ4UF4XXmla51NIsQfokWiNdF/ojUfNtdKU4xoxqgHt7ihUOUkGoOoQutQBWI4zdUBr6E6jcehcve0xIk8GFkgpPwIQQvwohPiXlLJNhSLj4+PbQcrJoUtotaVEbzPaNL9ukL6pcfRMtHLv+z+wq6SWm8do+apX7i5lRLYNvnkFvv0ngTHvR3WlqLIhT0krMOp12EwGVF89wpLYMGIGQGDSQW6KTSsm7A+QajdgcjdN86kZPF1tiTaRF/BpBj2lP+iD5zyUqTDkGw+O4nU6LRywaaierjkDW1euGT9TnDbyhmACKqt2caxrdPdiMIMtFSklvVPs1HkVUi1gVFwINaAdvzpfu+CE7hZAc6nEpWv9h1xEQgdJfcDScjKoEB6/2pAuIEhlnT8cDXS8tDUfCqDdoXhrorcHXU8dxXFp7UQ6U+8JH1lKuSFkxBv12eaih59//vmJSjlpdAmtqQNh7H0Rm9RLnmzIzxHEZNBz2QAbxS4PL6zazwur9iMVP+c5SuE7LdXoADWPZHvkbfn0s3vRM+kYoWRN4r9rvQoeRdWMWXyWFgpnSQRHD9Bphj431U6/tDgSrUYkzYxMm4vE8Li0hFe1JZoR9rsbXBp+N7iOYFJq6ZVgpI9D0M8hyUkwkmg1Yibo8jDHa6N3a7K2n7dG22YOGtWAH+ypEatEAVC8SL0JrzSSYDWS5TBjUmoQrqKgnjKIC/rzHdlatEtKf0g9Rdu/sZ9fqppPXjn+ausn6rxwuVzHcVBd8+4hXcd6Zo9LayfSmXrb9ZMQQvwC+FxKuauZ124FbgXIzMzklVdeAWDkyJGkpaVRXl7OK6+8Qk5ODhMnTgxn7DOZTMycOZOlS5eGF+JMmTKFffv2sWXLFkBzdcTHx4cNbN++fRk7dmw4H7TNZuO6665j0aJFVFZqE1LTp09n+/bt7NixA4Bx48ZhMBjCVU4GDBjA2WefzbvvvgtoKSqvvvpqFixYENY6Y8YMNmzYwJ49ewC44IILUBSF1atXAzBkyBCGDh3KwoULAUhOTmbq1KnMmzcvnPh+5syZfPXVV+HEP5MmTcLnqcPk3I/NX44pqSciYwgLlmiVSdLT05k8eTJz539A314jGDh1KdVlhRTrMvihOJ0B32zktL5ZlJWVsX79ekBbqLDwxqs4WFSCte4wA+vX0qMcQnlMe69/hHnnv8j8gjS+L5dMPCWRi0/rwbYtP7Bp0yaEEIwdO5aEhAQ+/vhj/ufcgSRXbcO8eym1GWcT6D0dV00NfmkEaUQxOTDUFGl3CFJFlSp1ARMGVU+VR1JW60UnBKc40jDXNHKthAxGE2MudXpE5f6G0DqhxxvfGxWw1OYjTDaEVHEEqhCe4Oo6nZGEhJ5IYdRCAD3VmuHVm5CObIS3Rou7Ntk1A68GCOjN6Im+CwnojKA34HK5MOtUzE2Nc20JWFNQpaSiWvtcDQaFJEsz4yRVweupw+WuJiEhAUVRwgu2bDYbJpOJqqoqTBYrJoMuYsVkst2I0aCjqqoKv1+bQE1MTMTn84W/T3a7HYPBEK5WYzKZcDgc4cx8oUUrTqcz/H9SUhIejwe3W/MdxcXFodPpwobJbDYTn9ALUZlHOG2BJYk6P9TXlDXbR+iutaZGG8lbLBZsNlv496fX60lOTqaysjKcPCw5OZn6+vpw9Ieqqng8nnAfVqsVi8USXkEZ6qOioiJcICMlJYXa2lq8Xu0C7nA4UFWV2traZvswGAzhTIuhxV2pqam4XC58vuBE7lE+JwCj0UhiYiKKooQLNKelpbXpc3K73aiqeky71xLtlmtFCDEe+AVwt5RHzxPaXK6V5cuXR5V+6qp0uNZti+HDmxsM2tj7YPTdUSXINh9yMuWlbyPySv9ydC4PXjo4YuIprHfFo/D137SNp0+Hku1QsiPcTs09H98vXsWS0Ciszu2Cg1/BlvmQ0AuGXUPguxfRO/fCKRMh4OfH5InkDBpOscuLPyBJsBpJNKkYfC6E0YzREgd6E3Vehf1lteGuHWYdmRYFs68KYTBqo2WJNpEbMhhCrxni6iaLlRzZmtF3HtAMsdBphrkxpjjNjeGuinL14MiKHC0n9tbuHqrzI6JPpM6IN7EfRpMZvU6nXRAqm8S4gxYxYk2O9H/766GsSUk6nVGLl2+FX9vj1xbguH0BHBYjcRZDuHzb8eJyuY6a57tFQjH6ilebnG48b9FBHLfWTuJE9J5orpV2+SSEEJcCY4C7gB5CiN5Syu/a0sfYsWPbQ8pJoUO1Og/CsrsjRqX+XZ9TPmAGthRzxJLy3SU1UcUBPthYwK1j+0assAvrzTqjoeG2D2D8A5A+BA5+Db3PQ/ez30QacYDdn8DSRqlB4zPR1xZCxlBY+WcAxM9HoVfcePxaXo7yWi+q3URAdZBht2DUaz5dpYnP1+VVqfHqGZDRG3PI7yslMvUUhK9eM4omG1QFjbjQgS1ZM+56Y0M+FMXbfNid363NJfibhl9KpNAhHNnaKF9n0FwsOp0WHWO0gNuJNNqRthQspkYuppZcDEZb9CSmwaIZ+Kp8QGoGMKl3qycnLUb9CfnDm+O4J+RCZfBaWUmpPYhNdraeE/aRCyHOAhYC5wArgY+AgUfdqRlaKovVFelQre4qLbIiyKHRT/Bw4p+58JXdXPvqWr7dXx6ehAwZdYfVwKQhmZzbN4WeSRZsTX78Yb29RsG5tzcsP9+7HM6/H371FUx+ETKHau38HjiyFfLXw+q/RuqTAegzFjbNbbRNxVJXSIq14etUVe8nw2GJMETN1dA0G5uErQlBeXW95ne2p2oG0pYC9jRt1OwOukicB8FfpxlgT1Xzhtxk1wy1iDaGHmmkRLHixQS1pYTvAAxmbYSfMgCR0BNd0+XuBjMk5nCkpJzM4ReRPOR8RI/T0Vvi2LWriUdR6LRRetogbeVn6iDu+d1DCCGwWq1kZma2uSTarFmzyMzUQkfPP//8Nu0LHLOKz/H07/V6GTJkSLvfpbal4lBbmD179gmdw5boKL2toT0mOzdJKeOklOcHH2dLKee2g7b/ThxZkJQLgLfnaJ4uGsy7W13U+QLsKHJxwxvr2VWsGfr+6XZuHJ3LdaN6czAYPvibiwYSb2lhxBeXBiN/BVNeg/95FibNgaTeuM0peNTgzZm3Bta9DKv+DAe/aTCCKf01d0xFXqPVmA2IgBeLoeH2wGbSY5I+bZIyOClqNurJSbZh0uvIiNOTm6AnN8mM4VjugqCvndriyMU1NcWasbelaKNye6O4d71Zc0V5qrVz2gjFnES5R0dJbYBD9Sb8jpzoUbJO13yYYNA49xg6huL8g8wPzqGoqspjjz3WTHuhjfDNcZRWVoXnhqZPn05xcTGzZs06+ntvwnPPPUdxcdM4/fbjePoPBAJUVVUdtaJPV+KJJ57o0HPYGXSZBUE2W/sn+ukoOlRrXDpMfQM+vJXi3Cv4ZEVk2Jc/INlXWsfgrATKa3zEW4w8+5+9jMxNxmE18sCSbdjNBs7rlxJe8JOdk0uBs560QCnmxTfAEa2aCTo9NVPmc+PXcUgp+H/n92OMowST4obqQqj/N0x4SLudrivVRr/uKkRCdpRsabBS7xcIIegRpyfFFEC4isFXAwYLMikXrzRi1AsGJEp01Yc1o+y2aP5ud6VmsC2OyDCu0IpJnSE6bh60GOe4dO21gF+74KiKFhkS8Guj8voKSO6HGvDjDuhw+vU4PdoEm0dR8epstD4SmwbjjObHDlVcnz9/Pn/84x/p379/s7s99dRTDBw4kO+//74tR2tXOiJEzmazcfDgwXbvOxZ+2IZjd9qRmxCqCdgdaBet5ftg01uanzlvNXgbIjZl1pkcmPoFzv6/4IycpKgVc0l2I4GAikEvWLGzhD9cOhiDXnCgvJYrz+xJgbOer/Zos+d7S2r4pDyFcU+uonjndw1GHEANELfqj9wwxMjGQ05mvfc9gSM7YNXjULxVM57OPNj4muYb/vpvsOxOxDfPoF7wcIO/WGdAxvcgziQYlCRI8RYgnHlarFx8VjAPSSE+RUEnVXTOAw0j61COEoQ2+Vh5gJTERnH6wXS2mh/cEn0ehdAWqhgsmkE3x2s5VDxV2vL92hLNR15XhteYwH6XoNLdZC5eHP/PICEhAbPZzP/93/8RCAR4/PHHm21XWVnJ66+/zr333nvcx2oPUlKaWX/QDhiNRvT69vXnd5TWjqIz9XYZQ75o0aLOltBqTlhr5UGYdyV8ciesfgLevhx2LQu//PXeMvbu3kaPH9/gvfR3+NekGi4boN0F3H/xILYVVDPt5e9YtbuM28f35y+f7WLNvgr2lNTy8ld5HKqo529f7OZgeS2/X7Kdf28vRkqJzhtdBEG4CrjQvJOL+tkZ1zce82bt1h+jDfqNhxWPABI2vaEZd4BD36D7YR4VUxbhmjIfjFZ0VQdJMAYwVh9AKMGRs7dGM6iWBISvBhlQCCg+VEe2tkLTHDTYAW+Da0Px4PfUNhIogsvzjZrrpFHssrSnIb21UHUYyvdobpxQf/76yEyAiheTnqj84zaTAUszvvvWEgrRu+2220hJSeGdd97h0KFDUe2efvpppk6dSlZWVtRrAAsXLmTy5Mn06dOHtLQ0MjIymD59Ort37262fUtUVVVx77330qdPH5KTk0lPT+fKK68M3wWEwv9ay5o1axgzZgxZWVkkJSUxY8aMiArxK1euJDMzE6vVihAiovixx+PhqaeeYty4cfTq1YukpCT69OnDPffc02zM9Y4dO5g6dSo5OTn06NGDfv36MXPmTFauXBnV9osvvmDChAkkJyeTlJTEaaedxl/+8pdwuF9j1q1bx/nnn098fDw5OTlMnz6dwsKmi9FOnLae2/akyxjyzjwJbeWEtR7Z0pCsKcSXf4CaEgqdbpz5u5i48RYy1j+Oedt7DFp5K385ZTfzbxnJugMV/PXz3Ww+XMWLq/ezt7QWX5NokPfWH+aMnCSOVHvYcFDTqko4YsqN1jLwYiwbXuCafh7q/SqqOZjjpN8E2LFE+7/HMDi8NnK/in34Kw6y0pnWEBUSUKIX9PjrwWBBNVix61XiXPvRuQq0UEGh16JQLEmRo+JQKI7iheoC1M3vIF+7EPlkfwLvTsdfuAV3Qn/y/YkUBBJxx+dq+9QHJ5tMzUQP2FPQ6w30SLCSnWjFYTGG/6/xKBwor6PY5WlzcYRQ/HJcXBx33XUXfr8/alReVVXFSy+9xOzZs1vsZ/bs2ej1ejZt2kRZWRnr1q2jqKiIc889N1wE+FhUVVUxevRoPvroI5YsWUJlZSU7duxAURTOO+88vvnmm3DMdms4ePAgjz76KAsWLKCoqIgvv/yS5cuXM2bMmLAhHj9+PMXFxUyfPj1q/+LiYu69914mTZpEXl4eTqeTefPmsWDBAi666KLwuQNtonDChAnYbDZ+/PFHjhw5wvvvv8+OHTt45JFHIvp9+eWX+fnPf87PfvYzioqKKC8vZ86cOcyZM4cpU6ZEFPvYuHEj48ePJxAIhKvb33PPPcycObPV56G1tOXctjddxpD/V9FcWlOvC1Q/5bVeThUHg5EUGtKWRp4uh+JqD//Z1bB8XEqiq8cAdpMBg06QqqvFbtIHt+l5YqsZ9YoXgzHYRhg6BeIyoWwXJtXLhvxafOfcGTSqssG4uoq0peVN8JhT6NfvFLClanHFLU0OAoG4LIyuwzSug4mnKpy5kIAvnBjLLw0gJaq7GveeVbDyMUTNEQQSfU0h+s9/R9mGxVR5FJzuAAdrdfjsmeGK95jsWtifzgAIzeVi0XJFmww6UuLM5KbaSY0zUe32k+/U6naWujwcKK/DqxzfD/KOO+7A4XDw5ptvUlDQkJDr2Wef5dJLL6VPn+hzGGLw4MH885//DFdfz83N5aWXXsLpdPL888+36vgPPvggP/74Iy+88ALDhw8HtIUpb775JkII7rrrrja9n/z8fJ5//nmys7U5kREjRvDwww+zf/9+nnzyyWPubzKZuPjii/n973+P0ajdCY0ePZoHH3yQ9evX88UXX4TbrlmzhtLSUqZOnRoO48vJyWHOnDnk5OSE2xUWFnLnnXdy7rnn8qc//QmLxYJer+eKK67g9ttv55NPPmHZsoa727vvvhu3283rr78enss455xzuPbaa9t0Lro6XcaQN3dF76q0VauUkgNltfyQX0WpywOZQ6LD5Ub9P4jPIj3ejEkXOardeu4zTPtMcLAierLPH5CkxUf2NWNUDpf0t9J/y9944fIeLL3Yz/Jzt/KXIQW4kwbBadPg3FlaHpNvn0Um9saUmsuCW8/F2vtsuGoeJObCiBu1DncshVG3Rvio5ZAr6V29gaGFiyLjo82ReWhkfA+qhIO6gGg+nWvAp13Y6iu00MLkvpisWthgvbCi/+ov6JTI961T3GRubAiL9AckXmHRElSBpseWooX9pQ8OloCLjv/2BVTKayOzFvoDKh7/MepP+t3ahbammARbw7lPTEzk9ttvx+fz8de/avpqamr45z//yf3333/ULj/99NMot0togcjWH76Heqc2F9BCXU5VVXnvvfewWCyMHz8+4rWUlBT69+/P5s2b25Qzu0+fPvTr1y9i28UXXwzAkiVLjrl/VlYWn376adT28PvaujW8LWRkH3zwQVauXImqqiQlJTFp0iTefvvtcLsPPvgAn88X1tGYUGrpkCEvKytjzZo15ObmMmDAgIi2HbGgrzMLS3SZqJXt27czevTozpbRKrZv307PQcP5em852wqrGN0vlXP6pZAeHz0Z5/UH+GTrEf6wdDtuf4CsBAsvzDiT4f/7MXz1V/wq5J31AGX6THqU19Mn1U5FzhkNCZ0c2XxemshZvY30SDBzyWmZfLqtIXRq0/5iXr/mVDYfrKC4xs+YHDOnJziJX/YrOOd2xnhWolv5x3B7uX8UXPAQrH9JG2kPugwx+ApGlv8Lhv8e9nwBC6/VYqkzToOJj0L+Bqg+Ale9o01IemsQh76BNX/XOv2fTzXj5nZqhZDNCSAV0JtQDHG4ahR0SBwGGzqlyeKcxi4VxQ2qgscfIM5mpdonyKotojmMTbbrDGYwNYk9OcbCGykhLV5LwyoEVNb5ohJVReGv1yaqg8m4hKuAxquy7rnnHv7xj3/w6quv8vvf/565c+cyfvx4Bg0adNRud+7cyVNPPcXatWspKyuLyD9e76qAqoPaE0e2FnLZhLKyMpxOJzqdjp49e0a97na7sdvtHDhwgF69enH22WdHuGx69erFhg0bIvbp0aNHVD+h0fn+/fuP+n5CfPHFFzz33HPs3LkTl8uFECK89D20XB20VB1PPPEEDz/8MBMmTCAzM5NJkyZx0003MWbMmHC70JzBk08+GRWDHwgEsNvt4WXyIY3NzUu0NFdxIng8nlg+8h07dnQbQ17tVfnr/O/DZdEWbijgl6Nz+d3FgzA3qZyyp6SGez/YEn5eVO3h/z7Ywvu/Opf4afNYsrWMB97bgaIWYzboePaaM5g0ZBieGR9h2PQGBn8NvbKz2V1bwdKNB5k8LJOhWQPZUlDFORkqE9R19K4t4fQtj2rG8/sizbVwyZMgDOiWPxShRxSsg+rD2mTmqF9D3gr48BaY9pa2EChknGtLtEfef2DiHPjxI/juWRg3W5ugbYwMNESHKG7N1aI3aXmrTWbiLQJvQMVvy8ZUU6BNhgq95vJwN5lvkBK3W/tBJJh12gWlNjrm1x/X8EOMMxswm83QTDy6VhBaoteJCOPoD6iUuLRaoqAVdejhsFBe68ViPMqNqqcmOmWtDFBcVY+KICkhiVtvvZW///3vPPLII3z44YcRLoTm2LZtG6NGjaJfv368++67nHbaaVoom78e0TS7oKso6q6nMUlJSUeNkQ4ZuaZGu7WEcpiI5txoTXjttde45ZZbuOqqq1i9enXYeK5atSrqrgHgt7/9LbfeeiuLFy9m8eLFzJs3j7feeotrrrmGd999N+KYjz32WKtj8EMXjo6mMwtLdBnXSnfCbUqKqm351rcHOVwRXYmnwBntDtlfVkdZrZe8KoXfL92BElyp6VVU7n1/C4eKSrD0HoFh8nPUXvYSPlXHLf1qeN3xCldvmMr0mrf47Tk2frlnFr13vaKtcKw+rP3IQbsFz1sNlgQqT53BoZ89gTv3ogYBepMWJ772OSjYCCNugpxztdeaZDREqkjnIS0yBFrIpS0iJyhdhVr74FyA1aSnotbLnooARfpsquL6oab010bxSiPXRhMDZdP5EGffFOWGkgYr/vMfJC3eTK9kGz2TbM0uKnL7AxRWudlfVsuRag/eRhOZHn8gbMRBM/gVdT5yU+xRF+PIgzfn2pD4lADltV7yyuu44+7fYDabeemllxg5ciTDhg1DVSV1XoVqt2ZUGqcIfvPNN3G73Tz44IMMGzYsHI8sm3WjyGbdK+np6eEEUo1HuiHKy8tZvnx5mybkmrsgFBVp37G+fftGvdaUF154AdDmCI41ApZSoqoqiYmJ3HTTTXz66aesX7+eESNGMH/+fFatWgUQvrNpaQJ4/fr17N27FyDsFgppbkzjyJufAl3GkI8bN66zJRwVKSWlNR6q633ExUcnxlGlVsigKemO6KXjGQ4zSVYTJS4PAVWSm2Ljz78YyrPXDOeRywdTVV2llXHT6Yiz28nVlXLOtzdh37cMqg6TsvlZen73EP7h17F77PMEDq6NOobqc/ONOpQr9/2ccct7cYdyB3snvAKmONxJA/Ccdw/yrF/CpD/DqNsgXivQwOg7IzvS6VGyzmpISFVfEZUqF51e278xehOqJQHpq8eCnwFpNgakGkmwGTFZbAiDqaEwhCVJ+6s3gc5AfHw8AVXVohpOuQjG3Ku9joC4DMRlT2MfcS09EqwkmQUm1ROVGtavqBwqr6OyzodX0XzhBU53ON+LEoj+rLxK4Ngh5eboz14KPS6v1m9AlSSnZYSjIh588EEAarx+9pfVUl6r6az1KuHMhiaT5r9vPOL0KSo/7m/GWOkMzeZ7EUKEJ/CaC4997LHH+N3vfkdiYmLUay2Rl5cXzsoZ4uOPPwbgF7/4xTH3b+59AREhiiHeeustLr300ohtgwcP5vrrrwcIZyucNm0aZrOZxYsXR0S9AJSWljJ27Niw+yUtLS0c2fLDDz9EtA1lOW1POrNOQZcx5AZD1/Dy1PsUNhyo5ION+azaXUp5jZcjVW7+sWIvF//9a6a9/B3xNgtZiZH+8ElDMiit8fDh5gJ+LGoYrQ/MdHDXBQ0r/cwGHU9ekEDG9tcYZncyonci9188iAKnm38s38f2w6XYDZI6bFC8DfweTtUXQH2kC8Kc9yX7bGdwxZI6KnMv1oxpI/YM/x2/nLc1PEG6fF8tD/2YReH0z3HmfY9l/hTE13+Df/0G5l8DzuCIu+/5MONDSi56nkOXLaBo2mf4rJkNE50b39AmSqe+CZc+A9PfBbMDvyEOf0Iu0hSPEpeFK74fB6sUDlUHqPX60fuqMVcfIK72EDZ8CJ1BK8TgDVadr6/URuTBFZP1vgB1gWDUySkXwbUL4daVyJu+xH3qNAI+txZDXp0P5buhfJeWp0aG7m4CUWGZdT4lvK25vC8OixHjsVbnGW2Q3E+LtjFY8FvTUGV0QYsXX3wRRVEYNWoU/oBKUZUHnRDh1LZGPeFQx2nTpmE0GpkzZ07Yr5uXX8hts0JRJsERusGiXURbKOk2Z84cBg8ezP3338+3334LaGlrX331VV599VWeeuqpo7+3JiQnJ3PLLbeER7QbN27kySefpF+/ftx3333H2BtmzJgBaJEjoXStW7du5U9/ar6A2Jdffsl7770XvmvIz8/nnXfeISsriwsuuADQfNvPPvsseXl53H333eEUt3l5eUyZMoULL7yQSy65JNznM888g9Vq5c477wzfYaxfv77NOW66Ou2WxrYtNJfG9pVXXuHWW2896VoaI6XkvXWHeWDp9vC268/tTYrdxDPL94a3mQ063rlxJP/afoQNB5xMGpKBKuEfK7Q2PRMtPH7l6fx7RzFV9T6uO6c3Fr1K9cFt9AocpO/3f0XUFqP2Gc/msa+xak85L3+Vx2s/tzPiwAvY81cTyBqB/vRpqI6eyNpi9K5C2PEhlO7UROhN/Gf8UnYrGXh9Cn0cAc6sW0OvzU/CObfxadwUfr1wB0154apTuWT5BVEXBi59CvxuPLkT+OKAwqMry6is9/GL09O5Yxj0rt6gxb6nD9HcMd+/DUg4/Rp2Dvg12X0HUeNR8CoBkmxG8pu4lPol6LDLOs3frTdpBRj0Rm3RTsAfrL2pGaiKigq8ehtV9X5yHRKLuwRdwItiSSJgTsLgLkfvCeq3JIFOBN+P0FLEGq1RKXNDnJIeh9VkQFUlVW4/R6rdBFSJ3WQgO8na6myDAb+P7F45eL0+XK5qklJSGT/pEh5+4u/0S4uL6MenBNh1pJqLRpyK3+fBWeXCYjGT4Ejg6Wee4dprr2XFihU88sgjbN26FYvFQmZWL26cdQ933TQDo9FIUlISb7w9j88++YhFixZRUlKC0WgkOTmZ559/Ppyrurq6mj//+c8sWrQIl8uFzWZj+PDhPPDAA4wcOZKysjLS0lquxTpr1qxw/+PGjeO+++7jkUceIT8/H4/HwyWXXMLf/va38EToypUrueaaa6iursbj8ZCamsr48eN5/32t6tSLL77ICy+8QF5eHsnJyZxxX1jcUwAAIABJREFUxhlMnDiRO+64A7vdTlxcHIWFhZSVlfH666+zbNky8vPzURQFs9nMJZdcwv333x8RgghaauYnnniCzZs3Y7FYwouV7r77bqzWyAyN69atY/bs2WzatImEhASGDx/OU089xaBBg8Ln8Omnnz7hkMRjndujcaJpbGOGvBGHKuq4+B9fU+9r8CNeM7IXy38spaxJmNrDlw/h+nN64w2oLNtSxH2LttInxc4t56RzVoaB335ezJbChjwpr1w9mInLRkX6hJNy2fI//+buD3czMVfPb4/8Bn2FVqRCOnpROGwWgdTBZH33EMaSLTDmN/jKDnAodQykDmCX2os7Fm4Ld3dqho3XL4Cs6u9ZY7+QGe9H3pon2oy8M2MQp713pmY8GzPut3iqy9madRVXfRhp5G8YkcEfSu7E4KuC4TOiMiLunPwlPQeeQb7TjcWoQwJ13khfe6pNT5bJ3ZADPHVAi2XCysvLwRJPUZUbgcBh0WHWgyL1JNv02Jy7IhcexWVolXqkqo1YLQkoAZUCpxuXp+F9pthN9EiwotM1dmEEUCUY9ULLN95GnFXVCJOVqno/ZqOORKsRqyny7lKVkrp6N/E1+yNXmwqdduFpJvVAcbWb0prI71ymw0K6o5k0BW3gRIzNyaY7aYXONeRdxrXSNM6zM6j3BSKMOGjpWNPio29lE61GdDqB1agnDg+P/c9A3pmoMnXX3QxcOomFvT/m2UkN/sgX1hRSnzsxspMew9BLhTizgVGJNWEjXt3vct4Y9DKTVudywcJaHk2cQ9GZ98Kaf7Cz341MWp7OzxdWs2pvJVee2ZDAamdJPTvcybDyMQbtfZmLT4k0lPdd2Jfnvj6Ce9gvI3UE3QTVA6ay19lkshP4cFsFZbmXgy0NCjc3vCCElpc84MNk0GHUC6Rs/kslmj5r5IxWAiq1wYlAjz+AxaIVcjbqdUgk1Z4A5fUqZqOe/CoviuX/s/fecVLV1///806vO9t77+zSe126gCiiEkSwa8RYovFjNDEaTWJNjBpjohBjF7Gg2AFFOtLr0rayvdfZmZ1+f3+8l50ddhUUFPx983o89vGAO7e85859n3ve57zO65ykadFVPQp0l/CrlApig/UkhBqIMGtJCjUQGaQLMOIgWuHp1MofZMRPXCfYoCE53EiMRd/LiINoTm1U+gKNOIgXz7e0fQs2aND0SOBqVUqC9N9L2qtP6HRn9iL4KfFzGiuc2/GeH4FpYMSIEed6CMQG6xmaGMyecr8mydbiJp76xUAWv7EbnwwqhcSlQ+JICNFRXnKU0OOfM+vou3gjclAmDIfGI+BsR7fnP8zMbmRc8vVsOd4hkqHh2VD0Sfe5G3Ov5/7PipkzJJFOl0vEuWWZ3QnX8ZdV/jG8vq+NuHF5LNa/irWlAZ+sAVlmxZ4qbp+Sjk6t6C5isVubwech7Mib/CVXw5WzZtOoCCdJ3Yps8lAfF0xl8g2EqUMILVqBb+QtKPBBRwManCRYeis7JobqMHSUQnslJIwEoDN6BLty72fFcS1XYMTjk4kK0lHWbCfGosPawyOXJAmLTgkdXTKnQbHdhtfTFT9u7WJzSJJEcqgenbuN+GAjdrd/xdhgdeDxyXgMusAHV6UV8XFTVIB3q1Ep0HxLPPls4XSVMBXKrnj/ya3kZF+fjTF0aiWpESYcbi+SJAx5X3H9H2u8PwR2l4e2Tjden0ywXo1Bo+r14vw++DkposK5He9545Gf6I15LmHRq3ny8oFckBOFQoKcmCD+c80wJmVF8v6vxvLoJbk8PX8QhXUd/PPLo4TvfxHTxj9B/RGUh1bA+idg1OLu82mOreSyrsK4mbnRfOjLo3zsI3hyf0Hp+L9xWE6hvNWFtdONOT4H68S/QFAcW+p7G5/3C31Yk2dS6QsN2H6spp1x6cJD1SgVZCjruj8LP/QyeevnMTemmQRlM+a67fRTVrLw3Uo+tSzEOudlpC3Pwhf3wqa/EfLe5Yw0NTI83u/Jq5USf8gLwVLwgWhMbIyAyBy+yXmQqz/rZOWhVjqcPkoabagUEqnhRlQKieQwI2FGDREmDWmhWvRKr5ADCMvwd5tvr0JytKJV+FdBsixT1ebA47Sj8Nipa3d0/3l8MhqlAqXcw7NVagSTJDRVGHLF6cW4zxZOW3dHpQXLSYU6xgihqd5S1jvUhXgRBenVmHXqs2LE4cfTNLK7PJQ02GiwOmm2uShptNHh7L26+z74Oekvwbkd73njkZ8vyIgy848FQ2iyOTFrVVgMwqgmhxrYU9bCw+/tx+2V+esUE4Ztrwce7GgLrFTUBmEyGnlgdgJfHq5jSGIEK5UzMcdcRGywHp1awU3jU3htaxlLN5Uwf9gE5k0fT0qTEghUh8sOU9KZcwVPfRCYRBwW5mK+egvrEyYQGxNDv5bPRY9PtQEOLIfMWUhfPUxk7X4igSyVlrgpr7Nw9TEWzGkUlYknIPtQb3qSv85YwrEWGatbQVJEELnaBkFTVOkgfgSuIdezZFlhwDhkWabd4SHa4veIA0MBXdt9XsF3twvvXAmEa4Jw6SK7NcJdHh9egxado5FoUzS1HWK7UpKIs2hAAp8qAYXsE4bb6xaNJE4Bl8eLLINapUBxGgUtZxWSJLoFqbRCFVLq6v95Qn7A4zztFnDnIzqcnl7MnXqrE6NW1UuG+X84+zhvDPn51GRVr1ESf1KLL68sY3N6cXfxjx0eSRg2ly3w4B4Gon3CQ6woksmIcjE6LYwVuyqpbBUTVyHB0/MH89Sagu79X99WQZA+jbhgPUlhBsq6CozMWhXXjE0l3+VDq/LLmqaFaZmuP0ropkeZPfI2pP3VSEc/9I9j+l9AHyK87hPwOMnKf5pp6X9EdvQus1ba67F6VNS41QyItZAZbcKkj4MkIcKELCN7fPRBmadX2KAveJzdRrz7mq52gk1htHTJgJi1SlSuJpSeTix6N/pQPT5ELFvrbBIJU7VeeOLt1eK6CpWQxlVphbxBj9/B65NptbuobXPgk2WCDRoig7TfXfhzmvguDW6fLOPx+lBIkihYUigEO6ejro+9T7p3Xo/YdpaN+9nWDD+BvjgTZ0qk+LHG+mPhXI73vDHkCxYsONdD+E5E0EpKsN/bXnrQzQVD7iJ6u7+9lxyajjNqOL7Jj9AZlsNvNktsPF7PiORQNCpFtxEHCDVq2F3e0us6nxyo4dXL4wmemU1RfQcen4yMzL83VfCfMU28O8VNgZSEor2KTMdOYrYIFTrtvtdg6DX+E8kybHsBpjzQ6xqatmKy45TIUQOEwes54Ub8kkH9Mhl08kEdDVC8Fva+iTZ6IL8ct5idZf7xS0jf3mKuJ05O+HXhxJ01alTE6D0orR3YzCkcb5fRKD3oVB5cXg2humCUmlZhxK01geftqBft3SSlaNrchU6Xh6oe977F7kKllIgO0p1Wqfl34YRa4clwuL3UW520dbrRqhTEWnQYtSoklVaMvUdfVlR6f2zf5xOftVeLKlJjhJD67avp81kc75nCpFVRj4Tc44UUYdaekTf+Y431x8K5HO95Y8iXL19+/hrzkg3w4WKG5NxERvhQChs7qWxx8ETtCB6c8waWqg24LClYQ/ujPLqK8NKPKZ79EcPSZRaOVDJEX8tHNYHKaB1OD0G63rc/yaIgZvtfCBlwDZ8cV/J5ofDKV10ZiuqT24mzNxE36fewKVDzWjaGIzlEwq+y/y3Y1GHENH1DUB+ViG0ZlzM2yQDNe/EueBfl138WrdxG3gyD++DSyjLu3W+gXtelC318E+NSi3nl6qd4c2ctBo2SlHADBk0Pj8TtEA0eJKUwUsqu76pQIasNSD072ys16HRaMvUa1AoJhasNjzmOKhskmiUMbuGdu70WfL5QlCGp4AyUSADE9RTBQhqghyE/mYkEgo0UbtTi6Sp316gUP4i50tzc3GsCe30yNW0OrF3UR4fby/EmO+mRXfxyS3xXB6M20JkFF/6E5+22Q0up/2TWGvGyNZ1UOfsD0dd4zwYMGiWpEUYaO5x4fTJhJi0m7Zl5qD/WWH8s/NDxng0K+HljyPvqGHJeoKkE3lkETisJe57i4Tmb2FErY3d5yIgycdio5qFjQdS1O7C5bFyQPoObJl/JDctKupM9Y1NDuWp0cIDz63D7yI4ykRCio6IrpqBVKfh1fw+6Lz9GF53FnxL1VNpGcKC6nShXpb9xgtMqknvN/vJp98QHkHe+whdDX+GPmzpo7/QwLGEhj7vUZE7+A2z7Nzja6Zj4ENaYCWSVfYTGa2OddiJ5C98TCcTghD5vQUNlIRFbnw7YZixZxeShVzDp2kspKipCq+xRiu3sgJbjeDRBQkNcQjBVlGocqHHqYjGqWlG52vGoTWAIRaVUiaWpxwktZXiCkgjTSZhtx7u9eLXHgc/nElrjfcXElRqx70l19uo+EoValYIWu4vadnHvg3RqYoJ13zvc0pd2idvr6zbiJ+CTZZxurzDkKq0wzMbI3to1rt5FTNgaQR/mfxmeAX6s5geSJGHUqrpf5me60oFz26jhh+CHjtftdp9xWOa8MeTnLTqbu+Pg9qRp/G1DDQerO7goy8ylUfV01nVwcZqFf+0U9D+X0shT2wMz9ltLmsnLiuR3M7N5d1cFzTYXlw+NJ9exm7em68lvM+P0ymQpa+m35UFh7WUfEbue4YGZH7O3PRa19oh/TNtfgNG3gcaIT2XkgJTBi3tNXD3qWe7qkYTcXWHlL9tMvJg3BEPefRwwj+dvW1o5sKGBaSmT+NWoUP6xpoJEh5U0xyFInQKxg7sNhizLHK+qwWX3EtGHCImMmLA6nY6Ojg4sZiOSuxOvs4M2fRINnTIKSSLKAGa3E4VSjdXpo6bNi05lwagOxu6W0doUJISqurjm4k2nclkxqk29QjEKRwt4o0SM3BwtWB8gjLcxoouGGB1wjFGjRK9R0tnlmSskiXCTluNN/vxGu8ONsVNJhPnM45wKCZQKqZf2jlIhdasxKiSpb2qeoo8pqVR/i1jZ+YezYcD/X0N7e/sZ67ScN4b8hC7DTw6fF6r3ivCJSgMpeRAziAMVrXx6sIa6Nom5M9Yw8viLSAoV1+Rq6D/WQFLhq2g/FeJEg0NSGTjjGW78wkZSqIGjddaAZCVAU4eLd3dVMDEzgssGhDDQ4iD0rV+DSkviwPmw9Z/+Mal0XYlUK5WNrSzZ1shlV8ZgyJiJInGkv32aIYpFu1K5aVIG80bJVFp7Nw3YVNpB3cz+KCsLuWZNTbfi3weHrZRaJRaOTqZWY6PGkMmanR2Ut+7mF8MTGJcezv6KVj7ZWcqMRJmEUXdh2NCj5ZYhFF9Uf5QIcaKS0uOoXHoMRhNWhYWqdr8BbnaqkVRqJIcblUJCqZBweHw4uneRcHu9aBQSKDXIxnBUtkYkXV8PtyT+vC7Ro9McLV58So3wdHWW7lJ/l8eHT5ZRKyWSQ410ur3IsoxWrezVUAKgrdNDmEn+XoyWvpbSGpWSGIueyhb/72/WqVEpFNS2O2i1u9GrlUQGaTGcXECkMYFCHdiEwxxz1miVP6dQxc9prPD9xivLMm63m/b2dlpaWnpJEHxfnDeGfOfOnUyaNOmnv3DFdnjtYr/npzZw5Op9rCtow6BR4vTC9SvrWXrZb7hAf5TLVl8Lw6+HI36FOWVLCcNr3uKmETdyVcRR4mwf0hEezZHRM7ltnUyL3Y1Fr6at083H+6uZFdmM1tOGOzgFde1eIfk64zHIXyE0utOngbUO5/QnWX1MQ7PNzr46N1PTp8IXv8UTnErhoHuocKRx/bh4YoJ05PiK2NTUOzyVEKLF2JTPYf0A2joDP99b0c7M/m7SIyK45pXd3auI9QWN/GlODolyDY+4nkT39QbImYt71jMoStbSbEhBzrmUyAhRjavWaHH4JA6UNaFQtuLy+nB5hDeqVytpB4rcJ7xhQUtss7u702JmnYrq415CtDJqpQKPDCqfC1/jcSRHK5LXX/0o6yxIrcVCW8V9klyrMQLUemTZh8sr4/T4kGVRdBSkV6PuUSnZ4fTQag8MfwTpVLibvh9LxOFw9FnRJ8syPq9grSgVEjalgl2lLjp7dB8qliDSrO0twev1gdfrf0G1Vpw1j/zbxns+4uc0Vvj+41Uqld0NobXaU9NnvwvnjSEvKCj46Q25zwvbXuw24r7QDI5N+Ae//6SIfRWisnJkSijXj0vGbmuDdfcIz8jaW8vYUrOV26ZdRsiKGwEIAcaql/H0xDfZ2pHAliJBuUsJ1dLffRDjrs+oGPl7Ej67SjRtqNyJd/bTSLZGFLUHYP/baNUG/jD0PpzOFFKDZBSf/hmMEXw55J/cvrodr68JaGLRqERGRGoY6itnbnYSK4+KkIFaKfHYODVRq66l6JLNwOGAMWuUCvRqJSWNnb2KN74+XM2LwW+hK98gNhxeifrYZzRd8Sklcjz94v3Jtxabk3ePdLJ8Vw3DkkII0qlYd0w0MPi/CzL5ew+KJcCAWAtZ0Sa+yK/lsqHxWB0eVu6r5rrBQfyx6T7Kh99P7LBZ/OHDgww0SkxQlRBkLaA+dCTG+H4kBmvghYv9OQMQoZX5byBX7kQqWU993FTWqvJ4YruLO6dmsPlQA88tGIKpi1lT1mTjseX72Nv1O6dHmHjx6mGkR/obA7i9PmpaHaiVEjHBgUJMJ/Duig/pN2oyde0OYoP1pEea+yzeKayzctUzG3tt/881w5ne7+wkMk8H54Om0eni5zRWOLfjPW8M+TmB7OvmNJePfZQPPWNpqdCyr+J49y47SpsZkhBMgtoqWplJCpG4Ownu5EmEHHz1pI12xisPY0mMIEShZl6KkWHyURIO/5fO1Jncs9PEDeOWM9xYj6RU0+IMJqajGMOOpeJ4p5XEjXfz26mvEaX0gdNK5ZC7+d16e0D89a3t5URNz8TlS+Fh43ssuGwhrT4DyWaZzPZtENGPMJOOvIxwNhY2EmvRMT4jgv5xQSglesm9AkyMk9Dv/yxwo9eNrm43o2JsUFcmdFb0Flrsbt7bI6ROO10erhyRwKbCRjw+4RWfjEM1bdwzMxOLQcPqQ7XdzTe+qZHpjMrC4qqhwepk5b5q3vPKGDQphBqzqGlz8IRZIjEmBNKmwMH3/CcdcjWs/QtS41EAImv2cWn8Dvam3ceLG4qZkRtNvdXZbciTwoy8dO1wCus78Plk0iNNAYJUVS12XtxQwjfFDUxM0jIiI57xWTHdx4No41djTOfe57cAYrXx1C8GcemQuF6xYrVSgVal6HU/9N/Vjeh/+B9OE+eNIT+hN/yTQqmGkTfj1ATzbNNIajpcON2948wFdR1cmxEpEmzuTmgsgJy5cHglAJ7wfpSlLSJ9/W29jlXLbhI0dlwmK161iWrdeN5s609CSAT9FV5u+bKUDxfmMPjjGYSmTxMNkU9ClnU7ytBREDOINlVEQGebE3B5fbx+TGLugGRGH30SStZ1f+ab/QxavZkIs5ZnrxjMsdp2PthbxeHqNm6akIpZq8SsVQXoo2REB+GtyUVZviXgOpIxAt66XCz7x90FefeiVipQShJeZAbGB/PKluP83wWZNNlcpIT1VjickB5Oq93FfzeXBmy/IEnCcHwnqtw5oFGREKKnpNHeJWYmjH2wQS1i4ePvhpqDQodckiBuKOx5LeB8uspNzJhwG+/m+4gM0mLSBj7uYSYtYaa+l7Qf7KlisLGJu1JXEVaxhjbPCFzmWyB1aPc+RQ0dPLuh3H+fZbj/w4MMtthINXshPKs7tp0YauCuaRk8ucpf0DUsKYSs6J+2GcE5mWc/ED+nscK5He954w54PGemy/CDkTqFmjEPs/JAA0X1HYxJsfTaZWyyGfOhZchTHhIT8+D7YG+k87LXqL5sJf9OfAZPpxUGzA88UK2nJHQ813xmZ8FnLhatbObRjS0ojWE89FkR3xQ38a8FA8necrd4QdgaRGLrJCjMUUirfg+Tfk+Uyk5SaGAcTqWQUCskosxa3KGZAUYcQLHuL8Qom2mzu9hV1sILG0qoa3eSX93O3e/uo7jRxm1T0rl8aBwTMyN44tJcBhS+gDTlDyL51gVPymQ0TUf9HMotz0JTAQmhBm7OSwHgeJONYKOaJ1cd4+3t5azYU8FNE1LQdoUbsqKMXD8qmg0FjVw0MKY79Ds+2ci8kEKk9MnYQ/sRYtTw8Jz+qHowO0anhtI/ruv3icqB6z6FG9fA4k1Cw6UPeFESbtIwPCn0tCVgW+0uOtqbuej444TtfxGaS7AcfYeQlVdBm1/SoKnD1aui0eH20Vq6F5bkwTF/B3mFQmLhyCReuW4Ed03L4NkrBvOPBYOJ6KNh94+JczbPfgB+TmOFczve88Yj37BhA1lZWT/6dQrqrGwpaqSpw8X49HAGJwaj9Ngx69RcNcDInOBCtieEsqtCaIlPyghlpqkYnXkA5ZETODYhizBvA82KUJ5bq8CnVPGLgWZS9z8EeiNMeRCObwZDKPLABXxyxMORen9V4aHqdkanhhFsUFNQZ2WAsRVd9jTImio4xMGJULa1u9+lbI5FCkmEKX+A2oOEByfw3MXx3PlpNceb7AQb1NySl8a7O8u5akwy5Q1H6aVqbG9G5+vkzmmZXPtKYNNdnyyKY/6zqYQYi54JaSFkmhzsiJxPiDeD7Ju2YW4+CJKEKv992PL3XudWKiSuG5dCVrSZLw/XMXtgLIer22mxu9la3Ex8sI6Xrx6Iwt5MVvkyKuoH8cEeM8OTQnh4dhaDIiT6KSvRtsigzyWoagO46xiXNIyVt42juL6DIL2KnBgLUT20XDBFiL+ucZAyCUrXd39szbyUjyv0PH5ZBiOSAwuyvgt6tZIZMZ1o938DIcmCEthUhNReJVZjXeJXcV16OY4eCcxwk4YYV5cI1sd3QPQgCBGMBItBzeTsSCZnR572WM42fqp5djbwcxornNvxnjeG/KdAYZ2VBUu30WwTLIjn1xWx9OphTJGqeWRqJGOC22mvb2F0WibjMqORJEgL11OmiqKorRazW8fNqzsBE3CCSeHgbxfGo9lbDdUVcHwTxA6Bumps7c3srA8iPkRPTZujO65dXN9BfIiexTle4j672l/JZ4mHkTcjT/szKNX4FCqUJ6qIPv2NoNwBgwYuYMW4sRQZh9GmCqOgroN5wxN4cUMx8WNi6adQBfCv5eQ8JK+LRF8VIQZ19/c/gVSzl+enqNCpvDiDQ7hi2aEuTZmdXDMylv8bloDFZISKHYE3VGsWhg4IN2m5eFAcFw8S+uhZ0WaKqxsxN+0no+4lQt5+s/uw0ryXAdhV1sKusha+/s04tPlviVL0tX9B2fUSU467i/5599A/Lo5TwhAKc56D4q+RK7bjShxPS+go7jHHkhzedwOLb4NWrSQ5Jpw9097h02ojLq/ExYM7GbLvj2iU4kXSaHWyo7SJ+2Zm8/zXRTTZXMRYdPxjokTMhq42Yp0tog4h5MyoZf/D/3AqnDeGPDc398c5sdsutCwkid1lLQFGLMKsQd1ShCfIxejgdpq1CSxrtfDqDmFYp+dEUVTfwacHBEtlZo6GiZnhbCjwiz4F6VVUu/QkDrgW0+ZHhCdWsUN4sGHJ/Dt7H566Y9SFjeSVimjePWwnO8bMqnwboxxbUfQsx26rhOZSpPgR0FSEQmsRxv3LB7uNOAAHlhMWlkbYl5fxdO57vLC5qlvM60/btWTOeJWkbx5Aaj0OKROR0ibD8quwjL2N348dwi8/tnWHBOItWobbNpK89XfUDbuHizfTfS6A13dUMzvEy6iSf8OlS2HV76Bmn+hbOec5CEvr87YnhxlJDtHBpnfgiN+Iu6KGsL41AhArnhm5USQoGsHRDvuWda9EABG6yZoFiaNP77cOSRLf1xSFtr2aRFUzBCed3rEn4bgcw/zPy/H4BGXzrQPw1mXPMTZSNJ7+6mgdv/8wn5wYM3+Y3Q+VQiJJa2PQh5P91ZmWhD5DZecSP9o8+xHwcxornNvxnjeGvH///mf3hI3FcOAdKPgC0qdSnrM4QDhJp1bw2mQnOV9fBR4HOsCQeyV65dXd+wyMt/D3NQXo1UqMWiWljVYeuSCGS3OCeG5THdeMSSTUqMZqs1OdvoB0TzOK/W8j60Pwzvgr2lV3o6sTfTNDeZnfDrsb/ehLUCokLhkYRXjDvpNHLZKd/S6BzibRENkQAgOvgM2BJfL4vHgnP8h4Xyf/7MFgqWpzslMxiMT5byAdEbRGvnpYfLj7NfJGBfHeJdHst4cTrHIx1LmT5G0PAtCmj+vVXgygUREBFd9A7QG45iMRy9eHgDH8u38DhRJG3AjRA0S4KDIHOTSLa2sKuPFSL61BWexzxFDl0ZAcktQnrROrUAr0+WRq2wUV8Fvjyk3F8MalorcoiCTolcshc+Z3j7MPfJZfh6fHfZVlePmwzKhhoThdHt74pow5g2IJNqh5cGU+GpWCX05IIaX/NQTt+bd4Ac/7L5h/Omrh6eCsz7MfET+nscK5He9ZMeSSJEUDjwCDZFn+Qa1+3nnnnbPHwbS34Ft5K4rKbQDIna28Z5uBPigClULC45O5aoCJrJ13IodnIauNKKp3Yjz0NosuupgaVxwHq1rxeGVun5KO1yszPbqDAZVvo171CcOjcsm+9Cmue7eE2nZh+AbFmXlu8hzidKHsD5qMuqaMgXWBzY8j9v+LB+YOwq6NwuizoTCMhaI1AfvIWbOQ8t+HfW/626P1uxgyLoBC/77lIaN4s8TI7bkO/n3lIJZsOk6H08ucQbFo1UrcDQVoNz0VeF/q8tG0FDJ802KGqw0w80lYd1938jKqeRcD4q7kYJW/cEiSINHgggn3CGM8ajHogzltGMKEV501C+qPoH19DtqOevFRzHBKx/6X5zY1MDNlMtMjspEajvqPlSQISaK6tZMdpU0cb7KjVSnIjjYzKjWsd1Vk1R6/EQfxvb4YXqCsAAAgAElEQVR8GBJGf78xI7TLe2+TkWUZtUJBeoSJuBA9/14vpIBtLi9/XV1A5qJbmTbsF4Kiao7udY5zjbM6z35k/JzGCud2vGeLtTIe+IiTWzOeIzjqCrqNOIBt0p/Z1yBh8Tbz9LxcRqWEMjZJx97hT3KX4TGu8T3A6gkr6EiajqqzkV1lzQxKCGZEcggf7K6krLae/vsfRb37Jeiow9dey5u7aqhtdzI0MYR7LshiQlYU+5wxNCZeSKT1MGnBfdwKrwt1fT6Wt2ZS1+FmrToPz/BfCs9VoYSBC5CCE2H/Sd2SjnyCfMKr1AZRPvFZbl3nY+mORsqkeIoa7AQbNGRGmfjPxhLufGcfNnUf3nJwovCmQYScvnkOFr4H/efhyZ5DbeJsfj0lnexowVQJ0qt47oJgsrbdh9xWCbmXn9b99/lkyppsFNd34HT3MIgFq4XULIAxnA2jlnLDskN8sKeK21dWsHfoY8gnQiEaI1zyb4joR35VG39ddYxnvyrkyVXHeG1rGflVvdUPqzTJrJ+wjI0TllE7/F7B+bfVBza8PgVOGPCLBsX2Kqa8blwyKqUCtUrBzRNTuou8euLLY82CCnk+GHGPs0vX/H/4/zvOikcuy/L7kiRNOpNznE1dhQ5XVz8aUyRMfRiDr4MX4ldj2vkS6CwMznuaAm0uV77dgLsr9ry5GP55wZ2Ey51UtnRS2VJFRZOdAfHBXBjTimbz193nd4RksavWR2q4keHJITy1xs8Nvn5oMJdlZrOtwcEvDKGCTXECWRcKRgtgLv6YxyoW8ILuYp5fdDURNetR1ucLEag+ZC0d5kTWjXuPRreWf33jpq5rJeBwudl5vJV5SZ1kShXcFqNiR2cc61rDuWj0r9Fue06cQK2HUbfAusf8J20qgqA4mPdfGtscvPzlMWyuGl6c4MDe0oTZ00TC7hdEIwR7LeT93ynvfZvdzfKd5TzzVQFOj4+5g2O5e3oWCaGGALXGgmmv8srOhu6v6vL6WPi5i/9c+jZZigoiEzIgNJUmm9CoqW7zx87XFzQwtV8kI1P8TZgL6qzcsLKdyq5Wp+nho1g67m+kKhvEc2BrFInow59CVD/Ivggis7uPL2no4MO9VWwsaGB6ThQXD4zlzRtH8cqW4zg9XuYPj0elkChrspEUZiQ7KoiUcCP7KwNfKMnhJs45Oluh+GvY/qJYEY25DRLHgEL5s9Iv+TmNFf4f0SOXJOlm4GaA6Oholi4V1YsjR44kIiKC5uZmli5dSmJiIhdccAEvvfQSABqNhuuuu46VK1dSXy+8ucsvv5yioiL2798PwNixYzGbzaxevRqA/mMmoUu7CFPyUFj/OIqB8zFt64oxO9tJ/OpmPhz8aUBSD+CFgz4mpEUBgiu8s6yFX09NR69ziSKULs/OWLud2Wk34tTHsHRDScA5XtnTSkZiDo9sOkbM9P8ypHo5xsb9kDxeFCAdFdWSHnUQTo+XXRWdPL03hGvj+5F74I+QfTFE5kC9v5xeDk7mqCeOJ/fWUtbslzlNizCQoGzmr+NVRH9wtZC3BbIsyRRMe4XfHJ3O/LxxRKk6SEtKQL3hSaQeMqmO/ov4cuth6pq2ctNNN3Fxoo8qORRX+xFyNt8R8L184Vm0uLWs6Prdvu13WnuglMe/8Cswfri3mnCdRKr9CHNzJqPf8xqYItnfYcHlCZRsdbh9fF3mwhOsZeWOtahUG5g1+yLyq3pryFS12HnnnXdIzhmCZAyhqqUzQI+kqNHJl95hjNLVk2m3o9z2EtoTGu6HQN75X46MfprNhyoYOnYiD66t40iNuH/7K9vYUtTE9Rkubhwaw6t7W7nrnf14fTLhJg2PTY8hK9rM/CGRfHm4DluXquK8ofEMiNSwPb8AubOd0SOG8+qrr+JyCWfhpptuYs2aNZSXiyKi2bNn09DQwI4dgg00bNgw4uPj+eijjwCIjY3loosu4qWXXsLn86FQKLjpppv49NNPqa4WlbSXXHIJlZWV7N69G4DRo0eT48lH9fGv/DercDU1s17lk721aDRCUOz7zKfU1FTy8vJ49dVXAdFk+KqrruL999/v7lN5xRVXkJ+fz6FDIpw4ceJEVCoVa9euBSAzM5MRI0Z09+YNCgpiwYIFLF++vFvCetGiRezcuZOCAiHpMHXqVI4dO8aGDUImIjc3l/79+/POO+8AwnDOvGguG/cfw+b0YpYcTBk1mI0bN1JSIubljBkzsFqtbN26FYBBgwaRnp7OihUrAIiMjGTu3Lln5XdqbW1l6dKlp/U7nbB7n30m7MHp2r1vg3Q2RM0Bujzyp2RZHn6qfYcPHy7v2rUrYNubb77JVVdddVbGUlhnZf3OvVxn2Yu6br9gWTT1aGumC2bJsE94fG1FwHH944KIDtLx1RFx45QKiX9eOYT8ymZuiz6GseUQ1OyHoq8on/oi66WR/PGTI5yMJy8fQHGDjRW7K4kP0fDeTAnNJ7/yF5OotGyf+BZXfCYenMuGxjEp2s2c9TPAkog860nk/A9QlG3GEz+aHXHXcNMqO0uuHMAHO0vZXOFkYqKGm1MaiIsIR979KqaCDwLGYJv6BPvIJNRZSasyjJ2dMVw1wERo7Rao2gWpkyFlQq8QQIvNxfGSY/Tbeje6mi66oVoPV38ovLpT4K+rjnbHjU8gKczAx7eNw6LohIMrYPfLPB7xJJbQKP662r+akSR4bVEu4bYickZMgZJ1uLct5UHpNpYftAac88WrhhJm0nLrm3to6HCiUytYnJdGUb2VqtZODlS2MSM3mheuGiaSxv8e1Vtk64q3oN9F7ChtYv4SEYoLN2m6Kkm9fHTbWNYfa+CZrwL7k944PoVb8lKJCNLx1c58bCoLIVpYsrmcLSViSXBBdhh/nDOQ+NCfuLO60wovz4S6/MDtUx6EvHvO6jxr63ShUkgYtT9Or9FTjbXZ5uTpNQW8uV0Y3DCjhleuH8HA+O+XD/lO+LxdjUpOTWE9m/f22yBJ0u6+bOx5w1qx2+2n3uk0kRFlxj5wAK0lR4iwNUJQfKAhd7QyLkGDViX0L/48XkumVEFksIeVNaF8LYlCmTsmp3G0opErDLsxrnkYOlvwJU9AvvJ9wlurGRRhIjnMwPEecrVmrYqyJjvLd5Rzx9QMlmwoZk2lgdlj7kBqOIJDG85u3Wju6tKiUkiQHR2EV+0EQzhYq2lubecx1y9JSL2ODeUu9u7uIMSgprKukSe1L9OeE409JIs9nUl8XGRhaMKNDJElQgpXdI/D0HqMcSUvQstxALKG3k6V51ZCR9wAWTOg7rCQ743IhtCU7uNCjBqC+/enPeplpOYjaLx2pIgsiAwsM2qwOtAoFd3NqU+gL852bmwQeo0SVBYYcQNkzWJyHTz1VSn3zcxi3bEGtCoF84bF88bOSu4YES1WJMvmo/Z5uGnyJRxqjOFgjR1JgmtGJ5EdHcSil7bT0CVH63D7eO7rQj661EjWsRconr2Q2uCuYiF8om3ayZD9+uS5sUFcPDCW8mY7Zp0KnVqJUqHgUHXv1UB+VRu/++AgFw2KwV1eyPzxmSzd0cSWEv/LZs3RJiZmVrMozSVCSh4nhGdAzMDe4/g2dNSL++DsEMdGnEaxiaQQL96ToRJSBGdjnrXYXKw6VMvSjSWYtCrumpbB+PRwtOpvkdrtqIfmUjGusPSA7k3fhVON9WBlW7cRB2iyuXj88yO8dM0IjH103/reqDkA25dA7X7BHMuZ+63NV05nvD8mzhZrZSJwNRAjSdIDwN9lWe48xWE/KgYlhuDVjINtjwqJ2PJv/FzsyFz6133CuzP7EWLWk/j51d09FG8Iy2LklS/Spoqg2SkR0V5K0ro7u8+rOL4Jl8ZC/rAn2FnSwq8mpfHOzgr2lLeSHmli0ahE/rG2kHaHh398Vcj145JpUsAu9VCig82odQZs6jQyourJjVMwISOcdUfquG+MgYYL/kWwo4LgwhVY1Lfz7Dci/vpInonJ0k5iyjaiiO6PMaw/t2+PZluFHRAPz11jbuD28COoGkVIRrLEC3ncLoQefBnd4HlQb4Uv7oXSrjeJIUxQCqMHdO8rSRKWyASIPOmhlWWaOxy8t6eKlzYdJ8ig4r4Z2UzMjOiexKNTwsiNCeJQjbifQToVt0xMQ6NSdnWPV0JQDP2ULq4e42PZ9nJSw40MTQzm/d2VbCpsZHaWiYFSgShqUmrA6+bWMdG0+7TEBBvonxBCTZsjgE7aNTxq2zoZWLyKnOJVZCxaCSQKPvfo2wIpnPoQiBR0sbRIE/OGxfOnT/zhrAiTlgsHxDB7YAxrDgc2Sx6eHMqrW0pZe7Se5y7JxOd2sKay91TaUNjMItV+WPOA8OqCk+CSf4mV0KnQXg0f3S76pAKoDXD1B6deFWmMgmH09hWB21Imnvqap4m1R+v5/QcHu/9/42u7WH7zaEanhvXeuf4IvHstNHatvEbcBBN/56/I7WwDT6fomPQ9pXpr23rrIu0pb6XV4T5zQ95UDK/PEUVdALUHRbewWU92692fTzhbyc4NwIYzOcd11113NoYSAGVoCix6n462ZgxzX0DReEyUW3tc4HUyqPAZ4QX3aISraDpGbus6lAWrsCfm4TN3VRWmTISEkfiUOg4Zx/Di1moaOtzMHxbHhIwIbhyfwupDdTzxxdFuhbsOp4fc2CBm7l5MS1AWzSkXY2w5wOS2r0kauZAiTwhet5OHp0aT/unl0OxfNVyVN5P85HSyLDLzqh5DVyVifJRuQJOcR6Llt2zrERn69w4rc6bcROqOh5DH3olUvlWoO4JQKRw4H/3Hi8HRArmXCnrc/reFFOzuN2DUzdBwVMTxI3MCPQ+fDyp3wPalBNubmJJ8JTsjo/iqxMbNb+zm3cWjuxOPiWEGXr5uBEdq23F5fGREmUgxemD/ctj6nNBuyfst1tDRlDXZ0KgUNHQ4eejjw3R2MVwSYqPBK8JbZaMe5uptsdS0+3MRzy8cTFZUEBEmbbdHfgJRKn/XH/WelyFjsmAEjVosqlD3vSnK5odeA+H+YqY3vikLOE9Dh5ND1W3kZUTw6ynpLNlYgizDJUNiqW3r7I6Lt2FEYdAzKa6aXYGnYGySCdY96i9yai2D1ffDtZ+cmg5ZvRes1aLgqqlYhIXW/FEYc+0phLZSJ4prHP5EVLxmXdi9EjjTeWZ3eXj5JLEzgHXH6nsbco8LNj/rN+IAO18Sevvp0+H4RvGdrNUw7HoYdl3Ac3eqsfYVthqXFkao4bsNbZvdTUG9FWunm+RwI6kRfSSo64/4jfgJ7H1NJI7D0/s8749hw04X501oZePGjUybNu3snbBqNxzfgoyMCg0KhQwFa0S8HMQPEjcSir/sdaiyqRCs1Ri2PoU843EY/xvxRt74NxRqAzHD7kbvG8bBqk4OVrXxp9mZHG+y8UV+TUACVadW0Gp38/mg56lotpHg0zEkyEpI+RqyrNvJSBiF4uB7sG8HJI2DAfNg499A9pFYvIwXpvwFjdeG7p2tgeM7vpFZE27l3R5hUJfXR33sdDRXjKHd7iJn7+v+DwfMgy//6OeGbl8ivr8pUix7o7Lhv9P8D25YhiikOfHAVu+BV2eDz4MCyChdz/2T/8WWinA63V52lDYHMEiiLLpATZQD78GHi/3/f2selhu34nApuHUAeNsqmR0bwt93eciJteBqrIDs/pA1m8PKbGq6+moqJLjngixW7q3mUNUR7pqWwSOfHaHT7UUhwX3jLGQWPue/Ts+2aeZoGHYtDF7Uq/el2+PD6uhN0+t0ewk3a7lzWiZzBsey7mg97+2upKDOn6TVSx4ISWN2fy9flrrZXyNWSKOSg5kc0d7bGNQe6Cqo6tuQuz0+OpxuinxZvBv0FHoVXDKkgyE77kHRcERUwJ7KkKv1otNVSl6vj850nqkUEuGm3oayT+PZ2QolX/feXn9UVLy+NU/EoAE2PSX+PfWP0NUE+1Rj7R8XxF3TMvjn10V4fTIp4QZ+OzNbhPG+BU0dTh77/Cgr9oh8lUGj5LXrRzIi5STGibKPuL9S852dms66DfseOG8M+Yks81lBc4lQKCzdiFSXT7dJmfIAtJWLyaWzIFuSkLRGYTx7IiwNDiwHQGqvQnbZkIq+Ep+57URve4TFk15GY0ojKcyA3SuRHann9inp/HNtER6faC929/RMyppsvNCD2TIny8QjmjCCvnoIRc5cwXOuPSj+InNg2PXsD53BO9XhDGgJZ6SqgL6mrVoZuAydmBnO0QYnCz4vRadWsHLu82Ruvx+Ftcrf17Injn4GqZNErLz8m0CD01Qoml3kzhXc88I1vXpnJh9eyvSMR/n4cDuhxu/wgNx22PavXpvNJZ/z67gMdJ/cIpQf1XqmzniO9pRRfLHibcYMvxkuegb30TZAJBsvHBDD5/k13SyWZ9cWcsvEVHJigojVOUj/8ga0tYIVgCSJqtKT0UcD48ggHTdOSOGJL/zFSCqFRP9YobSoVEikR5pptDr522p/k4yhicEYHA0gpZOSmsHLiywUN3aiUGtJiw4lpHxNr2sRkiLCWX1gf0UrSzYWc6y2gwkZ4WhNwby+rYxlByTenfVXhjZ/Jho2nwHOdJ5pVEp+NSmdLcVN3dpBQToVeZkR/p1aykTNQHOpCAUd/ijwJBGZUJvvN+InsOslGHkzWGJPa6wWvYZbJ6Uxq380NqeXxFAD4ebv7rSTX9XWbcQB7C4vD39yiGU3jQrM90TlQnimEEk7gQn3fKfkw1m1Yd8T540hPytwWEUcvKVM9G48OXO/fQnkXgb2ZuRjXyAhweT7hfzsoRVCk2Xcr0WS4wSbx96EVNq7s4vGHE5xvuAeA6RHGLliZCK3T0knOkhHdVsnErB0U+Ay9ONjHVx7wUUMK/wAjnwEeb8VFZMA9YexTfozVy73sjjPzJ8/PcK8HAN/iB2Nrtpf4ORIyMMXmsakzBYK6zsYkxbGZUPiWPjSdgAuHRLHv455CAl9guFZXqbpjtFrEWoMFx7TwAVwUNC5UCj9k6suX/CuJ/8BHK29vr9PocLrk5iSHcHwpFDcTqdopaY6yZORVH0brvBMdCu7jDiAuxPLqjuwLN7k38ccRXaiAaOmBJvLS0q4sVv3BqDB6uSZrwp55boR5CbHw6xHhSyD7BPJqfjTLzK+bGgcOrWSN745TrRFxx1TMhgQFyhpPCo1jI9uH0dRfQdGjZKcWAsfv/M6TBJx67DwSMJ61mHFD4eRi2HHEvH/0HSY9TiseVDc06HXipWY1khRvZVFL23v7tRU3NDB3MFx5MYGcai6ndUN4QyddE/v+3sOMCI5hPduGcOO0mYMGiUjU0LJjg4SH3a2wWf3+CuWL3hEhIlO5GsGL4K44cJ5OBnGiO6k7OlCo1KSdeLap4GTQ3EAh2vasTo9gYbcEg9Xvg3F64Szkz61i4v/LTWULeVMz42A9hoI+un1dc4bQz5jxowffrDbKeRLv34UnG0w/Ma+M/edzbSlX4pBcqKu3ArRQ0TiTxcM47rUBfM/FOXkXfAGJaAMz4L2Kv95tEFsbQvjQKU/KFrUYKO0oYNZORHsrbLy/q5K5g2P79VJHcDu63nbAz3rVp8Wu8uKV5bpdHt5Y7+VrLz7mZywg6i6zbhTp/GudQAPLStgaGIII1NC2VvewiWDhRdj1qqw6NW8vUME0F8HnpmayVxzDO1xeRyPmIJSASkpGRhNZghOwm6IIb9dT0FnEOEaNwNtW4kNUsOGJ6FimlgGq3QBgla2UXdyhSGT5TsquOq/25mUYuCXWZ1khKjEA38iIaTSwLg7RYHKiZi9xiQm7Ml0QI8DOmoDnoXMKDNv/XI0L28uxaxToVEqenU0MmlVwtNOGiP+fgAizTquG5vMpUNi0SiVfS7PJUmiX0wQ/WL8huM7n1tztAgVDLhcvDTVBnjzUiGsBnD0U1iwDLJnU1DX0avd3qcHqrk5L5VD1e141KYAdtEPxRnNsy6olAqGJoYwNLEPaeCmokDZibV/FvmItKlCdyY8U4SGYgeLEF5TF7VTkoTRN/pf+mdjrCcjsY+4+uSsSMKMfbxAwtLF33fB5xN6Th/dRkpnC2yPhcv/C0ljz9KITw/njSG3Wq2n3unbUL0blvVo6vDlgzDvNRHn8vq76bhH3YGidj/qDQ8Kz3Pi7/zVhtV7uvfzjbkdW78F1IeNhPAs0sLSBPe6Kynqy5jB3treGfND1e00djjZWtzMlSMT0SqVZEebOVrr/25hRg0pnq7lWs5c4fV2wZ5yAYecUczpJ3FhkkTMJdlsL2vnL1treVo7mAFxeUxSRlDqsQE29pS3sKe8hXunp5EYaiBIryI9wtyr2vB3GxzELPyc5zbVsHWf8K5nD5C4/8IE4lRaPnMP47dfHAbE9xuRkMfzg48T5fOC1wnbXhCeecMRQYXLno07cSK/W7Kb6lZxH9494CS/Ts/yQfsIAsGUCE0Tol+JY+GGVVC8votBkScoaGpDoDFX6cAUjbUu8FkYnBDM0/MH4fR48ckEhEBmD4gmI+rsVVNa9N+PkXDK51ZrgoRR4t/rHg94HgGRDEybjObkBsyATq3E7ZVRSDBrQB9entcj6KWyVyz51aduUnFG8+y0cJLj4nWJBOeA+RA3zL89JBkWvSdyWZ3NED0QYocGHPpjjDU31sKfL8nlsc+P4HD76B8bxH2niKt/JxoL4L3r/Iy49mrx/5vX99kS8sfCeWPIt27d+sPVw0o39d62cynMf0NMlIajkDkDVdxAzCtu9IdNuultgfrclcp4vkm5H7VSYoauGLm+Gumip0VBj8eJwtbAxOBOPjnpkoMTQ/j8YA0dTg//2VTC05dm8cSFiby0q5lNhY0Mjg/iNyONxB58mZYpf0UVOxB93W58Qcm0hA+jImQUY0ztpGc2sWQHrDzaQWKIgacvz8HYUQqmIO5fXcrjFyYzK1RNrUNNrLqD3Nq/Y1I9xOs3jGTJ+mLCzFq+KfY3Jrbo1Wwoc7K11B8i+exgDROzIpiQHs4jXwQWvOyssHK0XwRRhlCIzMEZ3h9H1TEsrUcEHzo4kYJmT7cRP4GxsRKGtgJ44yFxj+NHwNwXBAc6YZTfoIH4/NIlIgnqtosV1NwXICydrR++1OtZUCkVqJQKFo1MJCPSRFF9ByFGDekRJozac/cYf6/nVupjWS4pwO2gX4yJtHAjxY1+1s3145JpsblY9svRDEo4KTna0SBK8Lf+QzzHA6+ESb8/pfb5Gc2z00FYeq8GHySM6tuzDU35zlXGjzFWo1bFVaOSmJARjt3pJT5E36sW4nuhraKX/aCjThj0/xcN+RlB38cSL3WyiL+mTBBsBa8brw9Ueb+FQyvFm/TwSqE/stXPdGjqfz1/2i6ztiSfNVeGYvzqPug3RyRu1j3avd+4wSYuz53OikPC+x2TGopWpaChhwzs1uJGRo8NpsHq4OVrhtHh9PCntYVI0u3QCouMCeTXa9lQ2h9Pkcy945wMOPQcz7pu5uMjwjsubOjgzvcO8d4vIhm0+9csu/AhTOVvEe6uhmNf+OPXQ+cxOHMGT/1iMEdr2xkWZyBe20mtS8e+GgdjLU1cfkEDPhQc8CTywKZONhc2MCY1jHaH8BInZkYwIjkUh8eLLVRH29VrKew08rzvXsrLXVyZezUXZxmJjstCV+U3OCCSg/MjKlFt9GuPU7kTdr0ilswnxxYlSag6RuZARw2yOZYyXxT1ZS1kDhuHLMtIkkRTh5PC+g48Xh9pkSYcLi8FdVY6XV6cHh9fH6ln4ajEwGRbX3B3ilht3SERi40b9p3FHadE/WGoO8K8HLV4NnoaJI9LrOCK14lcROxgaK0Q18v7LXzzvD83MOQqeG0OcQkjeemKu9la5aW0ycbYtDCGJYZ8u5Ep2yyYHiewf5nQjhl3Z9/7/1TQB8PFz4qw0bFVkDFdzB9j3wnecwGFQiLlbGnimCLFs9yzQl5rFrTPnxDnjSEfNGjQDz84ZYK4cScEqvrPEwnEdY8IT2/qQ7DpKVS2RhFuGX2bWOJX7xXe+iX/wu5wcsQRxptlFr4utfNInomU9h0QmStiusrACRWz71keSc7nxivvxWttYLvbwiOrA7PW6ZEmDrco2V7aQqfTwa3LDnbzjwHq2p1MzIxgcnYkXp9MvFxKbcR4Pl0bWE3o8clsaAqiMu1hphe8SXPKHD61h3NcsZBci5MhTV8QrBI5gfIWO7aqfC4oXYKxYgOOyCHMmnwvqncWdoeGMo3hBE15iXp9GBFmLbP6R1Pd6iDaogsQALttUhrflNSwp1y8LB7dYKNNGcLdyWrSI03MGRTLx/uFlkSEWUuc7aTkMoj44cR7e9HtjtVaOVLThkLSkxM7mOP1du58ezM2lxejRsnTXpm8WC//t8bJ+gKh2JgWYeR3s7J59qvCbq7+pMwIthQ1MjQxOKDDfS8c/hg+7CExGjcMrnjzh3lNVbvhtYvBZSMUYEe84HdHZOHzyRwqqeRYuRe9NJapRhu6N+f5axWMEXDRs6KfZ8ok2P0K1B2EuoOkFKwi5cY1YMk59RiKTqL1qbSg1Im8Rv1RyJkDyXm9DGjPeebyeClusNFodRITrCc13IhCEZiz+UEITYGxd4i/M8AZ2YSfCuGZMP0RWPMH8X+FEi76B4Sm/qTDOG8MeXr6KZIK34XIfnDdF1C+DVxWEZtdfqX4rP/lgq1i65Ic9bpF55kpD4j4YtpkWPV7to9cwvVfyoCdO0aamX/8AdT1QkSIQx/A3CW93rx6g5GcsjehchdMeIEgnYr2Lj5ydJCWyck6UnxlfLQgkmqnN8CIA1S1djIg3sLnB2rYVNTInEsM6NobCDOm9cquS5LEHatb+WjhYv62vZPNxfXdny3Ou5HISi1DVS3Y2hoYu+93qBqEeJGusxb2vBpQ9IStkRGdm7EnRqJrPsa9M7LZVdbMb98/EHDNF5/xK4oAACAASURBVDYU8+upGd2GHODlzaUsGplITLCeX09JJy8znOONHeRG6tB6sumFpPHQWCiYK2Hi4T5Q2crC//gZGsEGNYvzUrvvj83l5a5PKvlscg1XJCexviulMCAumCc+9xdcgVBCHJ0aitPj41t9rPZqWP37wG1VuwXl81SGvPagqD+wNUDWTMG4+OZf4OqxImmvFN53RBbbSxq5+tWjeHwyGZEmJiZ8cdK9bxDe+dQ/wfND/QlgEAn1hmPdPUGtnW6cHl/flLrok0r9R98Km/8ulvUgntlpfxIeeo+KyRPzzOn28u7uCh7++DBen4xWpeD5hUOZnvM9GmF01IuwgjnmO/nVPxRnZBN+Kqj1guaaPA53SxXq8BRh3H9inC098jPGCTWyH4zIbBh+nfACesYizdH+npg9YQiHwQthw1/B2U66p5AIk5gwecGNaE4Y8RPY/iLMeV48tJIkGj1E9oPwLMi5hAGr5/P1pTJbF+pZuyiM9y4LJbliJRUtTowaFf2Ce2t9aFUKwoxqNnXpWhtUPqIkKw+PD0xaDU20UN1Vjl7sDmNzcXPA569sKae508v8Jd8gK9TdRhwQpc/tlZwMS/sx4r5cDF/cS3LrNuK0zl7quT6ZXqwbo1aF1eHmaE07ty7bw73vH+D1beX85oOj1EWOh8wLu/eVQ5IhYxqsvAWWTsJevo8Gq5PlO8oDGBqtdjeNHU6uH5fcbXM63d7/j733jo+yzN7/38/UzEwmvXfSQyihh957FwRdRVARFcsq9l7WdV2769oRREFFFEGa9N5CLyEhvfeeTMpkyu+PO5nJJFjWhZXv7/W5Xq/8kSfPPLmnnefc51znuiiz6ImzZtmOBbpryKp0LOmAoKB5Ov8Kbc3ULJhJ3p1uNC2/0UwrTRaDUHtehuMfwZcz23Zxl7ueW5WFodnEmzvSbM5CfnoF2tr0rudWXAasVx5Jl6toNVvYd7mMmz47xtT3D/Lh3gxKazspXkSOAe8O+jdKnT2It+PA63ahtja0f88yyxt4YWOy7f1tMVl4ZN1Z8qt+h16I0QAX1sEnI+Df/WH3S1Bb+NuP+w/xX8eE/xWUGgjow8rDRYJ/fqVhomuM6yYjv6rwihI88uZakTW4hTjojgBYnP2ozz6F1iUEZVM1wSdeZdWsn1l51oBOUdr1mjoPrOe/Q4qeJLLL3EOw7zVMfRZSHzkd92GueO1/SrBgnH0onfQp/6wYzKoztUAxt/Sz8ubcHjz6vb388OTESHykWtv3WVebBlUZjPPK5IebepHZoMRNMmBSu7H0+wwAmk1d6YxGswW5THht7kyrY/iUt6ChGORqsetw8bcbJ7dlgMaoKTSV5+JOLay+gdBBz+Lt3MdhJ+Drou4SyG8dFEJhTSOXSw22Cce6JhGUH9tRycqb/o2y9BxS6UWkmlzYsBQG3sUF93G8+XMdMsU5DK1db2o5lY1oVXLuHRnBh/syUStkeFMLWJEQXIgzedUMj/TiQLqjoUOc/69MOlosojEYOkRkxn1uFXIBFWldA3tn5B0Tn6GO2PcPcY2fn3Q87hVNU/4ZCjs0gE/n15M3Yg5hBccdz+0+U9xg+y+2c8xB0Da9YzhfUMvtX5zAahV00vVnCpHLJO4e2cEf1TOCwulrqMs9j9HYQoT2CjsSi/mK2vYAZXUtdGbG1jWZqDQYhXb8r6HwNPyw2P774fdA7fq79Or/D9cG100g9/H57ybWHOAZIeqfB94UW95RT8P2J8WXUpJhHPUMa0uDWJ4zh8TAm7i9dxUxjWfpnvYR/4hOBL+eWPUBSPVFtkuaveORV6aJmmYH1Kp8qcYD9yPv27nmRgMHiuWsOGUvSXx5ophnvPQ8OSmW+hYTGqWcgxmVTJLSmN+3O9+eKqFO7gkZu1Bn7KIf0A9o9R/Ao07PA0JFUK2Q4aZVUtNop7EN7OZhU+lraDbCyc/teubxN2ANGog0/mUwtVLukUC9VYOb0sRZ+UCiG04QBASceYfPx6zkpdM6ThUY6BvixvTeAdQ1t3LvyAiqG40MCHMnvbgGTyeJ4isIFmWVG2hCheLkSmSpP9mOF7bquWNbc5vcrIElI8JJynbcVSQEu/HernRuHhhCsLuGxweq6XbxRVKHv4sVca5aIeOhcdE0Gs2czK1Gr1bw3PTu9AnpUH+vKxbDJsVnIaAP6Hzhqxl22t+ljTDlTbGb8v0Ns1xT1+dIXZHYjRnKRYlFqYH+d0DqZrwq3+EvCV/x9gExtGQwmllVEcNTI55GdexfQjJg9NOi3FRwHBorYNSTgnvtFgaxU8HZhwvncwBYMiIcCaHq56pRkl3RYGvSZZU3MG15Oo1GFaDi7bE6btC4O07oDnnQVqZpR/v3zN9NY7M9bIenToXPb0xGAuIG1xmnvxBaKVexqXlVY8L/AH/meq+bQD5r1qyre8FuI6CmAApPCKGkae9CVRatOj+eT4/g27MiQ8+thMN5Or69YQZBjakoUn6Co++JQY6cQ1B6kcb+S2lQeeIVNhRZTa4Q1AFwciPfeyT15WVEdhwY8o5la17XmuHetEqMJgsXCmu5t5+OhcE1qD2CeSDMm/7+KjJMDURFTUWVLsTmUetpHfM8I2qCmdLTRFKRkec3XmTZhGj2XS4nrbSesbE+DI/yJqWkjhAPLeM9KyCtg0Z6UD+kdQsxa3042P8DntpkpLi2isQQZx4ZJfFaTiRPD3iagBOv0mvXzXwRdQPV059h5blGXtmSgtliRaOUM7ybM4PNmcwq+xRpD5T3eY9VnZ7fzIQAdMYq5J18SHOUouYvk0All5Fe2sDi4d1Yd7IAhUxi/oBgTuVWY7JYuVRcx9fTtQSU7EU252MivHux7a8GzGYroZ5anJ2UrFg0gKKaJnRqhWP22FwP258RU7og6t+x07pyt1M3CybTb6ntBSeK4GsxURsxg4tBf6FQFUpAlQs9Ep/ALWKcmM49s9pW1phr2UbziFmsOpaPu05F94hwtsrCGXrbXLz0GiTXADA2wr5/Qv4xcX3XIEjdKhgQgX3RqxXc2C+YQ+nljIrxQe+kYNvFEioNRmb09ifYQ0dOhYHGDj2X5w424z3ucwbXbUNRdgESbhU3nE5sofbvWYS3jrfn9+aJ7y/Q1GrGTavkXzf3IcDtCoN0nXElGzv3sCsP4f0XuOox4Rrjz1zvdRPIv/jii6uvHqbzhJMrBFMg7ygkfUrK8I/59qxjNlhQ04xFpoBdz9vrjBvuge6zYeTjaHa+iLYyDTwjsQy4C0mupMEkI0fTnTJlKB6yAvGlCeovgoZCzeAGib2dyqPRvnqyKxpYMUnD4BP3I6vNBUnCPPAeDLJZJPipKAp8Gn2v2zE1GzDqAwneeAuDR75FmWtfApsULB4eTkOzCblM4rGJ0agVcpKL62hoNrHvchkzJnmDs79QlPONFyWV1ibSE55k8bZ6WwZ2LK+BV/cU4eOqZa9yGLeoXaClDn3BXvSuL+LspOhQPzXzUFQ5AVvt70+i5RH+Pu0N3tiTR0Oziam9/Anx0HK6zMpAv15IBUn2t0Fu4sZ+3Qjy0FJlMOLlrCK/qpGPb+3L1gslrDtZYCvpjI/z5XJRNsHjRHNSDfbx7za4aJTIJQlDqwmLxWpnWlSk2YM4iF6JsWtNHXPrL5YcHBDQBxZuovnSz3zSPIUPt1cC2UA2d48M5+EYGU7HP3Z8yKk3efT+m5jQYxBfJ+Xx0uYUGlpMqOQy1i5JpI8rYGkVQzAgOOBtmvHtZZy+oe6U1jUjl8GulFJbCWt/WjmZZQ38fXYPXDSOdViD0cxdO63sWvYyQS5qkF+h+VhfRv7FQwSHx6LwjGR6rwB6BrpSZWjF10VNkPvvNMEIHSI0Y9p7T3KV2Fn8Tp3x34trEhOuIf7M9V43gbzdZumqIrC/2F7mHrEZS6h1epaNj8ZotqCUSaw/U0huZSOujXldm0XGejj2ERIWwYUuTUZWlow5biYyQx2V9c2sTctkST89Fu9YZO1+mCpnFtzQh28v6chqG/AIcHXCy1lFTx93+qa+LoI4gNWK/PhHzJg1jA8LIlhxOBuzxYq7Vsv7c30IbqmjyihjwZcXqW9rEOrVCl6f24vCmibe2J5Mi8mCt17N0lERvJ1UxuzE7xhc8QP+slpBrwSyTZ6YOglfnSmo44FoH3bmtXCLV5Sgb87+GNyCmN2ngf2XyzhfWEf3ABci8z9xeKyu4ABzPd7HNO5RQjx0KBUS964+jUou4+iCv6H65kYbW8PXRUNFTgvrTtkbbwsHh5JT2UBWRYMtiA+J8GRyTz+2/7D7F99Sq9XKiZwq3tieRk6FgRv6BXLroFCRmZs76WjUFogSiiRzZIcMvv/3aXrIZBA6hCxFdz7692GHP316IIsZcf3o3qkER8KtyFyDWbXrkk2HB0QfY/2ZQvqEuov+TeJ9sLkD51uSIEzolHfz0hHp44zRbMHTCd4aq6PZJOOTCyZ+PFvIkhHhRPk68/jEGN7amWa74T4zNY5AN82VdxolF2DtAoKrswXDZMQTSIPupZuXC92u4NNNVbYoodQVisGuwH5iShVE6fK2DVB0TmiJ+3R30LMvr2+hsqEFD2cVPvrfnjb9JVyTmHAN8Weu97oJ5NcEOk8Y8wxU5UBlJi3R0zhsCOWdXWlYrWKI5aFx0WQWlaLhCg3OgD5i6GLCK8IcoC2Lk5/7Fs3kNxmlLGO49iDyA5eEXvng+9qoaQ2odz/LYyNXcLlWvMT1zSY+P5TNlttC0Ow+3OVfaRsLWH5IsiWK1Y2tPLclg/UDHuHHQlfqW+xNt/oWE4czKjiQXmGj4pXXt7DqSA6z+wQiabQc8ruNwS4VBLkdgrwjuCuMdCYpuWmVGFpMJEZ4QY+PQOMhBP+rsgg/v4aV0WqyE4ej9emG4oyjgBRAre8grGYTL25KxkUt483p4VwsqKbUJZ7gJfsxV6RjUjqTbQpn7+XTDo9dfTyP+0ZH4KSU89C4KPqEuJEQ7PabI/KXS+pZ8HkSLSYL/q5OKOUy9rVpYUd4RCHzjIbKDop1WQeETd3JL8DYAAPvEjXqmjzIPgBFZ0WGGTrkyiUDoMFo6ZLAW61Qb7SQO/xtwqoPizJJzxtFnVuhoqaptct1ahsa4MTnIjDGTRMB9+j7oPWiYexrlGvC0de34KVX0zPQFV9LGXdWrEZ/5HtQODG23zL+XtSfsvpmXtmSAlh5/6YEZDIJb72aOH8XzBYrFqtFGHkg3HyKyiuI2f8sivYM2mKGfa+K5mp4V6lbavLhm5uFJEM7Zvwb+i6w/+4eJn464UR2FcvWnSW/qolANw1v3dibxIjrZxjo/6+4bgL54sWLf/ukPwKFk6CZ7XqOjB6P88rOPNuX0mSx8tG+DNbN8eC0KZy+kZNQZfxse6hV5w095iFl7HHcipuNQh0w7zDyskvUxN9GjjYRpd6Hbv1Be/IDZOUpWBpreHeXvfn04qRuBJx6U2hK5DoG8yZtYJdgkV3ZTL1bLFn5XVmiBTVNeOhU5HWgi/m4ONFitvDIunNYrOChU/HZtBH0G/4osdX7mBM/ix+SRZYsSbB4WDhHMytYMLsntFu0tdTDlscgcxeegCeAS6DI1M9/K0oBKh1ovNglJfLiFnvQvHtdGl/PD8FP7wTKcOSe4ciB1st2zrvtJWzLInenlLE7pYzXZsYQ6qnDVaP61c9CelmDLYgvSAzlX3vSaW614KSU8fa8BCbd9DWy4x9B1l5hXjDgTmGR1m2keA9lMjBUwob7hLEBwInPhHPNhL9fUa/Ey1mFv6uTQ4PX10VNiLUIv/4TQTZZlGs60M5uHRTC3lTH5z03tBG2PgIqPdyxXdTpu8/kcrWVFzalciz7IEHuGl6d3ZNhkV54lW9BmfqdeHCbfPJjk1fx9IEsDmUICYZDGZWsvnMgCcHunMqt4vOD2ZQ3tLBoSBhDw5ypyLmIl6IVRe6hri9mTW7XYyCy945BHIR+UcQYcA38xfemoLqRJV+dpLqtEV9Y08SSr06y+YFhhHj+tudlZ1yzmHCN8Geu97rhke/YcQXt5quFrL1QmUmFSdOFcmUwmqm2aHjtaAPbQh4lfeQH1CU+Rs30FfzcmkBdv6Vdt+yAZGoCUzNZw9/h9uLZzNqpZ+r6Jl5smEVpz3sxBg9F7+nHw2Mj+cuAID6eF8NsfYoQikpc6jiIknALtdqumhNR3lp8nMzM7dHVhGBGTz8udfKTHBfrw4d7M23Pscpg5Kk9dVQ1tuKmNPFMeAYr5oXzzJQ4VizsT2K4B+/clEC3jj6bVdmQucv2q9k1jPSEx0lqCqDxtu0k37CblYO2smXMFlYeK6YzjuXWomx01D+P8HG2cfTbER/gQna5KDtJEvhq4Ugbn/7XPgvaNnGjG/oG2YI4CM/OZd+dJQd/mPw6LN4Dk/5h97mUJHvjrzzVHsTbcfJzB4emdpzIqWR7cgkvTO/OgDB35DKJxBBnlo9qJeDsexw/3kYttJiFxnbuUagvITHck08W9KVPsCv9g11YPkXPgEuviZtJS52NDlqLlsd/TOZYG4unoLqJxatOkl5chfLid13W41t92oGxBLDuRL5tyGr7pVJO59Ww9kgqmtOfErNxKr4XPr6yV6hLp6BcnQcpm7uWGEHc4Ds3jTuhsKbJFsTbUddsoqCTJd/vxTWNCdcAf+Z6r5tAnpeX99sn/VGUi6wxUF6NWuH4lL2cVajdA7h5YAhFFneeTevGM+XjGfqTjns3V7C52AVD79u7XFKKGo/FAt9WRnKm0J4Vf3e2nOMe02gY+RI+tRdZmv8Ir9Y8yoiaDbi6eYhx7PWLBZti5ONYblqLQd8Nr5pTPDNCBAoQQlffzHLFaeeTDM14g5dHueLlrMLbWc3fJgYSry7m7uGhdg66So5e2ZWfnVbeRI0mBE59gceex7HUFvL3rSm4alT0D/Ow1zCNBtEQLjotlA7jZmDxiKJ4wocoJPAr2k2FoZWl2+t4aWch354swlXTdfDBTWXtYhgQ5K7lizsGMLmHH97Oam7s48eshAC2XCjGSSnj9fGe5JbVkFosBnR+7bMQ5+9CfIALkoQtiLejudVCWX2zkLTVul952rAsxS6d2hFWa5dAVVDdyKncalYczmHpmtPo1AruHx3BcxND6Vm6EbrPIjc3FxqrxWj8J8Ng5ST4fAK66hQmxvvz9W09WB2xi3EH5qEp7sAnb0sOimuaOZfvyFU3mi2kljVh9e0qGFXpFEK/UHeWjopAo5TzwJhIegW7cTq3mofHR9tumPfEGdEeeEU8r/QdQn1Q16EYPvBuoQFju3AmfDUL1t4izE4UnXYmfW/rGvg7wV2rQtFpxF8uk37Teu2XcE1jwjXAn7ne66a0ck0ROQYuriP87Jv8e8JHPLK3kbomE956NS9Mi2PJV6eQSxIrxsFij29p1IdT2ms0G7NkeGkkTin60f+GL3E69QmSQo0UOQ5KkzFEzWB3StfgearWmYmyPcSm/CCyNECbnwTDlonhlNYmSPoUAMkrBuvov+F84l8sbKxj+IynKND2pKbJhDl/F9QX41a/ntuKDjKp163gEoCPZy3nmjyZ5V7OpAkNVJmcCKaAik463QAxPjrcm/MwBg7mcvxfeWYfDOzmTrh3hyzcaoVza2HLw/ZjsVMxTnqToI0LBWcaQK7ivTFfMWsTHMms5OkpsZzOq7btAFw1SmK7hZLa7Eq4yWyr0wLE+2h5d3oIdZZI3CrPUJGxj35jA3FTtNLsrGXuTzW8Pa+TtobJKLb4VTmidu8bT4CbKx/f2o+sigY0Khl6tZLatlF2J6Xs15trtQXw9XyInth1SCx8dBd9jPL6Zgqqm21CaPsul+OGgQCnTHHTKzjB5L73iZtDR2PnmlzY+SIFk1dwOKuOlJY59Bk3jyFl3+J9+l+C5RE8EBCTsi4ahW2oqh05VU1kxtxFZNY+W9PY6BXP7sZIVh3NpVeQK/+4oQcvb06hyiCabGqFjMcnxfC3zSm4muzql1jMsOcV6LeIau8BaD2DUPvHCg0h2z88ZN+RJH0G414U2XlNDvT+i6iP/4apRTdPLc9NieKFzfZy25OTYx0/a/+Ha4LrJpBPnTr12l08fDQMfQjZ8Y8Yf/wOfpz0LimKOPxcVNy+WrBBVk5SE5/0BEf7vc0/T1nJT2pgXm8vXHRa/rLiJBvGt9A7fpbQ3djxLGBFO+ZFhobqyCx3pLgl6KpRNJYLf8WOOPMlxM0U23gAnReSqQlnSwPkHEIFyD0jeHh1CpIEOwZ2GO4wlONz+h2hOx2SSHDIJJTNFej3P2Y7xTt4JE8Me5I3j9Rgtljxclbx5MQIzjV5IAtdyNECI7cOUjC9dwBuWpWduleVbRf9aUfqFlRxs+xBHMBsJCZzBYND7+dobj1fHs3llVk9qK+vQ65U02yWuOubSzSbLDwzNpCbo61o/LuLYHDoXdSpW/D27w1jn0efMIuK3Fy2lSpYvr+ZxcPDbb6Jts9C6mb44Q57f2LQvTD6aYI9XHBWK3h1Vk+OZlXi5SwmUOMDXDiRXcmrW1MYGe3D6FhvR0pdWaoIsqdWiiGxsmQoTcYUOYn0wFlUFxjpE2JCoxJfCw9nNcUdRuPVChkPhOXhvts+wei682GY/WkXHZ4qTQiPr7/EkSxRMvkCWDBoLs+O8kQdPhj8RTYc7KHl5Rk9eGjtWdtjJ3T35XxBDcsPNLHjtq24NGSSXtnKoXpf3joodi3nC2qxWKy2IA5izP5geoVQ4vR2dVyTsQFr6macus9DHXwFMaqONzVDudCm6TVf9EZcg36Xw70yew/zUt8lYdIdFLU64+/nR0y4j8MN/T/BNY0J1wB/5nqvm0BeXl5OYOCvb93+MFz8od8i6sKnUGlU8tSeOm6OzqWwSjBAFDKJSFM6l3o9xcItBhvX+pOjxdS0yojx1dMstYraYUabWbOzD5U1ddwcKyel3I0TuWKKc2yEjoHGY5jDRyA//qE418kV4mdjdQ9DaqzGqvXmQuJbHKz1xNfbl3ClnIiEJbgWHyG1GhvNsMC5J36dJTLjpkHKT6T63YpaG4w0ciV1ZiWh1iK6nf4Ht8ZcpPeCSaRUmujpKRHZeIra+lpSaoORJD+WxDSRb6jixUOZpJQamNc/mJF+El6d3XpAuC11gsZQgI9OlKdyKxuxWiExOoj7vzlDfrU96L28s4D+ilbiJJDveBpZXltzN3s/rL4B57v2EtV3BE5VTUwZIiPYXYOizVyhvLycQK0ZNj/s+NyPfwTxs7AEDuSboxm8vsuuodMjwAVfvZqXt4gm3a6UMsan+fD2/AT07aqI7TVyc6vQBwkeRN6w19la7k3K+Wa6B9TS0mpmdJwQjgp01TAu1ofdKaJp2T/EhbCsDubO7cjYJcb9y+wNwvSAGRw56TivsCapgLlLF6JVKTh9Mp+mVgsJwa5MjPfluyWJHMyoQCmXkVxUy+5L4n/urXAh1n8kN3xzBHDUhjF36vcAGFpM3DG0G68dy+Bvw18n6MhzYkLV2Rdp0D1o1i+ABRu6qvOFDXWUxbVawTP690v91hbChnvRGMrpXXAI263irr2g6/trj/xFXNOYcA3wZ673ugnkSUlJJCQk/PaJfxTnvkWTfYTyQW/TN1xDrHM2tc1mQIYVMEsK0kzemCwNDg/78Uwxdw7rxs9l0Mdfj8olkBqfQWzyWsx7Jxuxnq1k6cgwHhoVgtRYQbS8GL2mPwUWT7q5BCFp3YVP6InPkAzlWHrdTOW0Fdz1o5F7RgXyyZE80ssaGNRtNi9MfRRlgZ1r/cIpNe+OXU7k+beQmqowJSwk030YDb7z2JIlUdtkYvMFURN1VoexYs5WvNStLFh1kQ8nu+NXfJEaq4yAvE2E1efhPOYH/pZUR2VjGQnBbpzIq+WRdedYNi6SpRETUWRutz9xpRazZ3SXJkp9z0UUnJcR4a1jeu8AtieXUFzX7BDE25Gs6olrbTmheZ3ols21UJGOk2cEUb5ddVKSkpJI8B0stNZlcsoSHiBbl4BCMhPZqqC6uIT39jkyLi4W1XVR7tuZUkZ2cQW9urW56/jECc5zm3zBhailzP3OQItJBMiN54p4cUZ3ahqNuJmrUZxZzdgmeHTMJD48lI+h1YrJLZAu+aVrkOBWtwdyvR9N+jDgksNpFqtori9aecLWFFTIJFYvHkSfEGG03NGTFCDMS0eEt47RMd7svWzfHU2M9yXMs+sAzj0jI1h5KJuhUf5swp+gyePwppoYZTnuWxaLpmXJxa6BPGgATH9fDMUZDTDgLug9v8v1fxGNlY67t3bUFwF/LJBf85hwlfFnrve6CeTXFKYWclSRrHMdxrZthSwc4It/YAiBB97kvgF38sGJOpKJRKfomuK460QN9usz9fg49eP2SW9xpEzPcz/bdVT+tjWNR8ZHE+yiJ9DTlZeP1nBLZCs+U/6NwliD0/pFtnNlZ75EJ9Ny1/C7eGtHmk0F8Hh2NQ+sS+EfM2N5bkQzo9zK0Gsk6hTBlE7/mubWVl7eX82enRUsHeWDn5ucNUn2WmRDi4kX9pTzzjhnjo/PRdK28Hl2DB+dqmdy9DIWjvJi8TepNr/LbRdLeHpKHJdL6vn4QDaD5/+dHiYjmty9WLziSOn7AlUNfkSM+4CAk6+DsYHSXveQ6zqcmwc5s/9yOZ8fzKa+xUSwh7YLPU+SQOvkxOniFkIVauEs1BHqXxb2l8lkgtXjHUd6/IPcc9KXzApx7dHRJp4c1drFtxNAJkmEeWrJqbTvLqx5x0ATIWRf9f4w4324+AMYytlnCKXFVORwjU3nipnWwx9Sf4I9L+MD3Oe6ijlj/4oiaiwq8xJI+8n+fNR6YdnnFSWmexvKwMkFb5kePxcnSursr8nwKC8uFtQ6MDvaKbCf3dafv46LIr+6kXP5tagVMpaNj6ZngCt6jZK/zerBwfQKkrIrSQz3YlikoKXBFgAAIABJREFUF25aJe/M781bO9JoMpq5e2Q4bholI2K8eXVrRzs8f26LdmeQTe2xw+fcYhaDYGo99LtNqFWaW0VjU/4fhAe9L7gGC8ecdkgycP11x6L/w9XBdRPI+/Xr99sn/UEYTHJWlEbhrHNibJyVz44V08c7AN/oadx/6V0mTZhHg0zCNySaWJ98UsvsgeDZ8aGkF4ta9clqJxKdolmfk0vPQFf8XJ04nVtNpcFIUnYlD4xthYzdvGc6D1VxVDuNora0gEAQLka9bwKVDq3FTIKHqYvZbma5gRpDEwsr30XhMQYyT1EXcxcfnahhb0YNPQNdeXxiDEcyKhkc2dWBJKWkAW3JJbwOCWW+eyJnUhp7BwcKjPh7WboEv52XShgU7sH5/Fp+zm5luWUZo4c8gqenF8s25fHcJDn37fJmfvyH6FVWvj5tpFteDXP6OrOpQ+a44UwhT06O5e2daVQ3tqKSy1gyMpxKQwtrzpoZ0u9RfI/b3ZVMkRNR+MQ5rCWjrJ7dKWWkFNcxrNcYKq3OuM9dyTcHqsmssPcKzKYWgjK+YnrsSJuLEoCLRsFEr3JuH3CBfF08jx8CN42csOINUKOG6e8JFovGTWh1e8dilLqO77eaLGjlrfY+hiQh9Z6Hf9152LxamJYs2CgkbmUKavRRuAX0FsGvLAU2PQgWEzHePfnn7HWsPV1ESnE9A8Lcmd8/hF2pXal9JbUttJotRPno+fL2geRXN6FVyQn11NlYTEHuWm4eGMLNAx0DY2K4J89OjSOttN4m3fDeLkdWzpYLxczt1SYQpvWwi4VVZorGZspGMdsw4lExBPdH4OwLc5bDdwvEzUyphalv/7bC5K/gWsaEa4E/c73XTSAPCgr67ZP+IDIrGqhvha8PZKFWyLh5UAiV9Qb0VjXnI5bhZKzEWe+GpamOj2YGcrG8lcqqKuJ1NfQ6uYSpAT2Z+5c5/Fig4KmfLnPPqAg2nysiv6qRWX0CaTSaSHBvgSNvQfoOJIDM3bj7HYChzwra17CH4eBbIvvRehA7ZSBezioqGuzNKrVCRoi6EUX8dGiqpipwJA/vaeRCseBX51U1crGoluGRXlfUxRga6oxnsX3wwyVjIzeMmMehQicaW7tmsGaLFbkkMX9gMJvPFVNS18z2dFg4xIMYby1ahURds4nPTtlr5ZIEUd4a5DLJNtRjMApT5M33DSGjtIaTBQY2ny8mIdgNmUzOC4X9uXv0Knyas2hQ++MZPRgvrX3aL7+qkUUrT1DQVp7ZcLaIZfVm7hgaycH8Iw5rnhwqQ3f0TR5N1BOq78tPmSZ6eCm4Z4gf0fnfQO5B4rzS+GHaHIpzknE9/B349hAqmGHDhEZI4v2w81lGjr2HDyQcZgsWD++G0SKhcglFUZYCvW8Wbj6lbRrvRachdCjMXwNad1pK2wJzZYYYuW+TQVCUX2D4rpmEzd1MdasST2cVwR46jGYzH+1z5KsvGhqGTi3q+K5a1a96SFYZjGRXNKCUywhx1/DurjTWnrCX42L89V0MTACqm8zQbxGtvW5B6RkpBMbOrBZvqGuwaCznHhY17V/x0fxVhCTCXfvEWL/WQxi8/I4m6S/hWsaEa4E/c73XDY9848aN1+zahzMq+PFMISaLFYPRzPKD2eSbPfBtSGZQ1UYGeFtIOPIAkd+PISzpeaabdrLo9FwG7L0VdelpZGdWUYOedw4UMSHel+d+vMiulDJSS+r5/FA2RpOFicFtfN0OkErO4ywzYU58APa9Zreia6zCedMS3p/oOOizZEQ4ZUY1rxX0pKk4hdxGNReKHZuQuZWNhHhoiLHm8MJoLxsvPsrHmed61aLL2uJwvpO1kdK6FnoFutqyu3bMTAhkYrwvORUGhxLAkBAt70SdxdXatdk5Itqbb04W8tlt/egb4kaIh5bnp3dnck8/Aj2dSYzyJ7fSQHaFgc3ni7hpYDCFTSpmb1PyUPYg6sMmonJ3dOW5XFJvC+Lt+GBvBrWNxi517/QasHpEEnL0eZblLuWnsPW847mBnuf/Lhxy8pPgzFeotjxIqK4tQocOgY1LxcSiTC4mKm/7iQTner5eEMf4OB8Ghnnw8a19GRXjzeoTxVwMv0PosbgG24O47U04bKPq2T63DaVdeOiyysuEmnJJCHEn2ENQ8BJC3Ph0QT+ifJwJcHXixTFejJefES5K7ajMhOQf4cL3Dg3UzLIGbl+ZxJyPjjLj34f5ObmUdScdjSNajBbiAxxFxtQKGf7uOhj/Miu3t7Fjis9CzkE4vQqcXATdsKVOCI/9N3ANFNRKz8j/KojDtY0J1wJ/5nqvm4z8WsBkttBoNLHhTFGXv50rqKVvwlzCW9NwWr9QjN0rtUhmI1RnCRncDoG5wOQOiHJCfaeSyE/niniwTwhd5y9BVXERk3tEVzealnr8pUpenhGPyWLFaDbj4+zEbV+f4x+jXdCc/Q7VxDlAVyGeMZ7VRG6YTZQ+iJEj78Eg6QhQ5OFak+PI8nByI63VF42yCW9nJS9M786RjEoMRhOz+wRyIb+GZpOVMRF6lvSQ0CisNFkU9Dq2GKeSk7iHjuflSS/yxt5CDEYTE+L90Dsp+PxQNkMivFixaABWxCBIO9RKOY9MiCHIXYNWpcBktnJrYgjRvnqOZlZy/9dn8HRW8djEGIZEeKEyFGFu6focLVYrrRaY2z+I49lVnMoV5ZUmswzT8MdQbrofqTYP19ovsU59W4y+d0RNnt3gWZJEoK3KgOABInCFj0QBJAL9Y8KwYEUll1Nc08RH+zJZpZbzz2FfM9itliuKs0qdciB9gBiiadcwVziRPuJf7M10Je/kBcbG+dI/1B29RsmEcA2DEs5iri3C49z3wobQJ16wSZoq4ctZ0NA2HatyhoWbsAb04buT+ZwrsN9c86ubkMskLB3oKxvOFnLnsG58eyKfpOwqunnpuHNYNyJ9XcGp7ZlUZcPaW+3G3ek7ob5E1PrlKmEcYagE9xBhW/ZfBuT/w7XHdRPIAwL+gAnuL6DVbOFEdhXLD2Xh7awmyteZy6WOgTTO34VV5+t4NKgBrdkI0ZPAv7egkgH0vkVkYnWFNMXfhELlhEouQ3aFD7VaIadF5UZz5BScMrbajlv8etPi1QuV3Cq+IOYOAUuu4nyNiud3JRPsoeHRCTF8ekCYN5c3A2NfpFvlAeb3GM/ai/a1T+zuTZAxC0wtyKszCT9s55EX37gJfXURznm7MfklYBjyBF6l8K+5Mfxzezpz+wVRZTBitlp58adk7hkZQbi6mvHaZOSnVohsLHoy9JkH207inLuTW1y9CZ7/HCUNJrZeKCIAA1+NbSVWX4qbyhsUXcsAoZ46pvcO5Jblx6kyGPF1UTO1pz8rDucAUFLXzO1fnOD7hXH02zSBmITH8NCFO3CibxscSqC7BqVcxucL+5NTWIwy/wjd8lag3HMJ6+hnaVR7Y9AF4Swzoh35pCh7pHVg3qhdRJBM2QRAg1scecW1qORyQj21KNvoju20RxDeqAq5jPJ6I3f8DAt6e/Ns6CjUZedIH/ASaZYglBo9IYQQS4fttGekqBFvuBda6ske8hp/OehJeb3ItFcfz+OfN/Rk/sAQqMrE9cDzji9aWbJIIPKP24M4CKGvpE9pnvwee1LLiPDWARKZ5Q3sTinlLwNDWHU0F7VChslipbCmifgAFxYNDmVqTz/cNEp6Brrh7SKCeEBAgCgDNdc4/v+SC9BngbjhfdWmq61wEiWkqHFd3uP/GLUForHqGvS7/T2vZkz4X+DPXO91E8inTZt21a51Lr+GWz4/bktOn5kax4G0cpsxcpinltExPszqE4imoFI0Iv16iDHrdmQfgNHPYL68k5M9XyC7tJG35/cmo7SBIHeNQyngnuGhtBhqyevzOBrvRAKKdlHmM4Rj6iEUlgYQrihnyvBlwh/UagFJRsHQv/PeKbHAsroWQjw03JcgxynOguSixZJ1DO3lLTzSo5mR4ydzvlpB9xA/BqhycDJewWvSyZVtObCjbjET+9zDpGg9/t/PJHrE22yrDuKBMZE8tf6CQ/30vd3pJC3QI1//qH3HcO5rqC+GyHHk+E5gVW1vNq1PJsxTx0Mj/Bm2Z64IBEfkwkB4wJ1X1KHenlxiC8wjor3ZesFRf8VqhXNZxfRrbSQs6SW+HvMJ3xS4c65CYk7fQMZ397MFWjetigTTBThwj8gO+99JoyaAFqULHifeRpG9V1w0fBQMXCKmZn26i/JB3lEAcqat5fldLRxIP4RCJnH3yHDuHNYND52jBoyfqxOPTIjm+Y3J9At0YkFQGQpGcW7AK/zlu0IMxmagmW5eDSwfr2JKQC1U54J7qMj+feOhoZxLlX6U1190uPYbOy4zJs4H77ZAZvKKIy/2TlokNSElu9DJ1Ta5ZQeUp6JuLGHNFBVOF1aD1Ur2oGk8dqyVSd29eCCmBlnJOeQaV+TBETj7uRDTScMdALOJacMSxA2jM+QqIUW7crL9mKlZlKSW7BezGH8ETTVCbG3P38X1EpfCoLspsbqTUlxPc6uZaF89EZ5OYlegdLJJCVzNmPC/wJ+53usmkC9fvvyqqYftTCl1qDD8a1c6yyZE46NXo1bI8Xd1Yt2pfPaklvHcuFB8p36PVJ1DSNAQ1AVtzTVzK40mSBuzklUH8ghy12K2WtE5yXl1dg8uFdWTV2VgYDcPwl0s+GevZ/bpXtQ1xRHhPYD8M42UNzSwdJSZk/U63ENnIJswBBdjOSqPID65JCezUgx93Nw/CL/ivfQ5/JCoU456GtJEZu9z8VOmyFYwRecNngth62tCYz1ynH33AFgm/oNgdTT55y/h7+GBsqmUsulf8sAuK+eLUnlqUnSXJliLyYKsLq9r2SdrL6Ypb/N6cjRbL4tR74oGI0vW1vHTqAeIPPBXartNJckYz9YfUgjz1jMx3k9MWFqt+Lo4kVNhZ4TUNZlw1ykd6vAAKhdPdk3YTZ1ZSUFVPcP8G7h3aCA79hzAb9CtwuGp1SDKGO5hIpMbdA9k7EaX/CO6XvOgPYgDZO3DGpyINPkNiBgtdkDdRmLR+/H1eR0H0nMAQfn7YG8mfYLdGdfZNd7Uwtx4F0I9BhDZcJLATXdiDEzkY3l/2+s3tac/0X56kgxgMeYRef5hpNkfChlcj3DwCKe1sqvhdXOrRQybeURQM+offFnbi3/vrcdotjAyfAkvykPpFjNZ1K07ImIMss9G4RM5BvTecOwjeqd8zWfTvyfAcgHl2nl2vXWXQJj8BvjGOXLFq3OFxPLpLwTzpvtMYXvXjtFPC+eizmgoFSYYfzSQ5x2DbU/Yfz/8LiZnf/56vjfH24TCdCo5X8/1o/dPE4VL0qTXIHI8y7/48v8pBcSrGcP+U1w3gdxi6cqq+KNw6jQSXN9iYldKGSsX9ae+2cTCFUlcLKrjxv5BfHehji0XSgENN/Z4mr/22UPAGaGbYZJr+eRQAX6uGo5lVZJaYg94b8ztSU2jgop6I/7VF5DV5QO9qG1q5XRetcP/7x3sxs3ftmdnGqCSd+fEcqN3ATKPUHx0Fvy/u09osIAYI3cLszuwWEwiS27fkh75F6aJr2GNv5GmxgZSzYEEBA5ivI8Hw5yLcNpwG1JVFig1fJj4Aq+79mB0oJl3lXKaOhgfq+Qy5FeQbUWpodZnINt+dBQBajSaybAEEOnsw2afu3lmey3QBJSw6mgudwwNY/mhbJaOjGBivB8bzorexL7LZTw6MYZXt6bYbrB+LmpQaPg+tYKfL9qz9VsH1jI/Jgj2vwZH3hcNx/53CpbFzd8J3e/yVNHALD7Xde1FZ+GWtfbffeOpbTSy9WJXGdcz+TWOgTw/CQ6+jbYynZF9bwOF2Gm0uISRViBet9l9Aimrb+adnaIpqFa4s2rKgySWpjjomcf6uaDp9HovGRGOn4sTSBJnvGfw9s/29e/PquWrEyU8M3Yw8qlvw56/Ca56wl9EWaKpSjQ/+y0SN7XqHEIMF5GSfnA0zagrhOIzwhD5pjVCo8ZigaTldqPns2uEdvrclSJj9gwH/74iU+88SewZKaiFfxSdCAAAinOr8dL0QC6TeGKwM4N1RbjVVWEe/wryPS8L4a47dlzVmPC/wJ+53qvCWpEkaZwkSR9KkvSiJEkv/KGFyK4egWZsrI+DyqEkwT0jwlEp5ORUGLhYVIdOJcfbWc2m88VYrIKCtvZCHXtVo0VtVe2C5BPLtuQSvPVqhyAO8PrPafQP82B/ejkyqwmP9B+4b5Ajt1ujlBPsoeFAmqPrO8D3Z0roW7iGpkYDhXmZDoqBxvoKmsa9Kvwc2xE6FKrsW2LFwTf4sSaC/doJzNtqIa2yldqqcjQ/PyKCOEBrE0EHn+RvA4xEnfknb411tr0uKrmM+8dEUq/2h+BBDmtrHfEUaa1+PDQ2inn9gx0U7fRKKItbxFtJTcgkeHCgnm/HG/lwqIE+nq3UNrXy6rZUZJLE32bF46FToVbI0KnkfLskkeendeefc3pyS2IoVQYjO5IdSy6bLpQR0ZoqylytjcJM+OCboPeBYx8JsSsQJYgrmCdLoYO7HNOp5PQLde9yPMpNgpwjorFXmgxfzoC0baJ0tPN5qLgM3jHo8/cyN1rcRLt56TicYRekajFZeOm4lVq5o/FGrJ+er+7oz/Te/sQHuPD3WT24aWAwUluP5UJJ1+x364ViahqaRLnq3iNww2eQsRvOd7gxpW0XNEpAUiivPE1paobS84KVkrFbNDfPfuV4zoV1gkU14A5RktK4iHLUzI+E3jyIevbsTxxVE/9TeEV2OdTiEUtBnZl/jNRye9Yyeh64m+Bd94ogPuoJcSMpTUZ+Jbu66xhXM4b9p/ivM3JJkrTAx0C81WptkSTpB0mSxlqt1l/267oCruaWpGeQK+vuHsyOS6U0tJiY3MPP5rTe3tiK8tVzvqArvW5zvpJbRoqtYKPMGU+dEfMV7rRVjUbC5BUsTVDiogikIXwSNUaJRyZEk5RdhbtWxagYL0ZpshgWaeBBr0JqFJ58ne/B2uRG3NRWaoLHsb/ak2hPf1L7/0i0pp4ARS0fFkaQtKWOF0Z8R7y6DGetE9rMrULnxckVmmtp8unLtowmxveyEuKhJURZg6I4SzT8OkFnyEe6tJ5JBSfxnr+WfUVy5DL4/lQBq4+ZWDPtecITMpHXl9DqE8/KwiBe/VzoZUd461g2Ppr9aeVM7uHH2YY6ij3mMLeflVhVKdNSn0B1Xoyim3x68vqoV3hsr4EP9mWw7u7BTOzuh9lqxd9VNNviA1worG7i1S2XmNs/mHtHRaKQSbSaLSw/mM2wUC3ai990fVPLLomMsb1c0FAqbrh+vaDkPADWwIFI0ZPE301GkcnKlajcglkyIpzDGRU23v7gEC0Dyn+Aba9BzFRR327tJDNw5kvh+nTkfWZ6FFDaP4LWK0yUppc3Ue8UR0N1E/sulzHAuYygnB/oX3CIsO63UT94LB6+/g7uRw4a8G3o7avE+dSHMPoxMdmafdC+K2uHW7CoJYMY3hlwF+x52f53SQKtl3gu5W03xB5zhIxte0beDm2noTKFSgytBQ8SN1DXIDGx+d8gYhy4fgS1bbs7tZ6mvneR8WU1iTHpKKs70C5b6sRzDugLGnfuvPP/LfPlP7MMdDVKK4OBXKvV2j6DfRiYCvxHgXzz5s1XrVkgSRK9gt3oFdyVEBjurWNyDz9O5lQzqYcfhzIcs+UBQTrRuMo5iK9k5Zlx4ylskqGQSTYxLYC58XpC9j1EN/cQ5KZGzg94jeZcI1isjIjyZv2ZAtJLqpnW6xgeB0UTNRQIi19AbdRcbotoplwbz6lLNXx20F6KeWxCX47kFFBQ3cztPwO48P78nvgF+3NEdhs6BQx1q6RR4cG+jfUMiGzl0xtCiPxxomDetG27O8KoD0ah8+FS/5cob1Hy3Ylcm08mwLR1Jr6Z14e+hyexd+jXvLrb7jWZWW4gtaSeG/sF8uj3F2zHh0d5cX/wUVQVdj0RRdkFhoYex1PXlzBPLVartW0c3YqbRklyUR2vbUsls7yBZ6d25+G152yvqZ+LE0tGhpNdWofJIxJFW3C2wdkHq1KHUReEevLrovRw6F3h6jNsGVgtSP69xKh8TR4c/9Reax77AvG95vHj0qFkFpSgrs8lungznieXi79f3oI5ZkpXDRWVDuJmgXcsAcc+5hmtD8e7v877nU6b0N0bvasHz268SHfnRiKTHkJWJQKUV8ljOAePIH3Uh/SMCMZisVJQ3UiEtzNDw905nCXee3etkvtjG1Dv/Bf0ngMBCVR6JKD3jENV2cYlV6ix9r0dKXWzsK8L6CtKcFazcDnSekLCLXDmK/E6NLQ5FF38QbBPTi637/x84oUPZ2dIkii1/BqMjeIzJpOLAasrMJds8I6GRZsEK8bcCj7dkekjeGF6Mfq8nV3Pr84WaqUBfa9qTPhf4M9c79UI5D44yrLVtR1zgCRJS4AlAH5+fnz6qdDjHjhwIN7e3pw/f56ioiJCQkKYMGECy5eLL5lKpWLRokVs2LCBsjLxwZwzZw4ZGRmcOydqjEOGDEGv17N9u6CehYeHM2LECL744gsAtFott956K3v37qWgoICpgeGMio5GJpMI99KSVSG2uaHuaqZrztKan4mypoBm/4HEqVsxyjW8Oz+Bj/ZlkF/dxJw4Hbe7n0GVfgKKT5B5y1Ee3ZRNWttov5ezisXDwgkw5aI60kGnGnBN/orXbr4Z/baHOZj4CafzHLfGH+zLYkFiKEczSpkXLaeiWcIqyZi/vtI2geis1nD/6ACs1suEuauIMRwXRhUufjDmWfjpQVGWAKp63cXWEm/i5xzgli/OIZeyWDYhmlPpRcwKacRFMqDxicJSnQcKFZdru25nj2dXoneyf1SGRnoyJtaHekJxcfaxBwzAqyKJCJ/hzOrhxQsbk/nudAEyCd65sTePr79Ac6uFkdHerDme53BjLKlrxkkh46/jYykvnIt/xnYh3gSi9qxxJ6fP49y9Jpd7hk1i6OwE3Kw1yGrzkKmcuVDaStLRvYRHFaFz8+BY40S8E8aTqM4mcsfj1Cu9qKxXMarluJCxbXbcjTXJ9WjdQpF1sD9rGPwEDSVF+G26B6wWlEC/1hZem/4ar+zIpaHFxKAwN26IduJidhGbzhVz7wwrslOOI/JO+QeQKjNodFfw4fFKPjucj9li5eAdAeREVtBkURBhSSd0/6sgSdTVVPD9zyswhiQiC/snQ+OLUFubyZKC2Z/lzbyefbiUdBbX3JP0io+lxnMK3pMH49pciNOFryBkiGgaHnjDtoaqqioMk1ahM+Tj4RdMltmHy2fyaKw9w6gRw3/392lwXACxBWtRXv4JZHKaet/OGc1w4geN5uLFiyQni+GpkSNHolAo2L1b5HPR0dEMGDCANWvWAAfw8/NHFzcWLix3eK0s3WeRohnAke+309zczOXLl9m/fz8A8fHx9OjRg7VrRanJw8ODuXPnsnr1ahobxed90aJFHDhwgKwsUV6cOHEi9fX1HDkiSAy9e/cmMjKSH374AQAfHx9mzZrFF198YTNPXrx4MTt27LAZRUydOpXy8nKSksQutV+/fgQFBdkGgAICApg2bZothslkMhYvXszmzZspKhJ9opkzZ1JQUMCpU6cAe9zbskUM8P3euPdLkKydjSL/Q0iSNBZ42mq1jm37fRkQZLVal/3SY/r37289efKkw7FPP/2UJUuW/Fdr+SUUVovR9nP5tbg4KRgU7ombVklxTTP+1lLy83KwAlHNFwg49QZIMipv2QGtTWyp8MVTryG3spGJAY1oM7bgk70ReUVbluTkysrE7bz0syOla0bvAB6IbyFq/cQu6ymd8Q3azE2c7LaUv24ustEi2/HJ/FgSM97FNfUb0HpRPPhF7j3pw9ki+9b/9qFhhHloGONVRZAxG2nXS0KwyCuG0okfU5CbQYNMz6oMJ/ZkNbJ4WDd2XColr6qRefHOPO++HeeTH4iL6bxpmfw2zclbOB7zJBfLjVQ2GFl7Ih+TxcqMnt7kVhs5V1DLw+OjScqu5HBGJW5aJS8M0zE57XmcSsT72TDuDQojbyKluI6H1p5jRKiGJ+KryZGCuW+LuGnNTAjgTF6Ng98owLJx4Tw4JoZNmzYyPVYjeO0qLUZ9CPtLlLx41ExhrdhJ/HtGINP2TLCLV8VMhVkfklJYRWnOJYySijUZKlIqjHw7MIdweZnQgt/0oChbpG52+N9nxnxJVrMr01wyUDcUQrfhQhHw1Bdt+vOIbHXiq3BmDfnh82hSuOIt1eI+YD4pDRomv3eIH6dJ9Nl1c5f3PHPmRoJqTjDtdH/Sy8Xz3j9fTejJV6HwlP3E7rOEnENAAisPZ/NDUjZ/7W3GhyrcPP2obJLwDo1hfXItq4/n4eui5vGJsQyJ8BRlw6Y62HAPXO4w4atQw5ID4BPLZ599xqgZN7H8UDb7LpcxMMyDpaMjifO/Al3xSjj4Dux+0fHY3JXQ44bf9/iOaKwWjdd9r4q6ft+FYnfVJp17LWPCtcD/Yr2SJJ2yWq39Ox+/Ghn5USBUkiR1W3llKPDhf3qRmTNnXoWldIXBaOJoViWPf38eixXifJzoLssl2rUBFycvLPVljDi4yPFBkkROVTNhXs70saZSXe/KMFcFpc2uDMz5yR7EAVPgQC6WOdLqADLLG9B4R2L2iERelWH/g8adKk0wbgnzGZ6znp8TdVxy6sOyfSYUchnDIz3oU38A1+Qvxfn1xfjvuJunRn3N/A4DqipLCzcH1JHdrEPa8ay9blqbz6a0Rl45qAMsgAgaG88VMaG7L2uO53FTcBXO+z7o8CKVozr6HmtC/8nL34pSSZinlofGRbPuZC73xhj48LIzMsmNi4W1tmZfTWMrD++oIWTyQ/QruRV6zMW5x2Ri3Fz46nAma2boGeTZgqK5NwPqAAAgAElEQVSxhUB1AwODtSTlN3Iko5IZCQF8fsix/jtInYc5PY8JvQPh9HIsVguy0mSe0L/Oj8mOJh1rLhiYHDgQee5BceDyFozlmUTteJC4MsEQGtj9Ft53vZHzUgzhbloxL1CTC31uxWKoQJZ/DBRqSvot45M0PT9nGPBdOJFhiR00uJ06NDHDRkDqFii9SHBpB454QBSh4ROYkRDAhvx6eoYMQ5FnZ8nUx92E0tKM+vAbLB3y/3F3nuFN3Fnb/6l3W5KL3HsDg00xzfTe0yC9LOkJKZteN5tsQja9V1I2lYQQQoBAQu81VBswNu69Sy6SbMkq74cxFsKEkCdhH573vi4+MB7NjKTR/T9zzn3us5z71wnfyf4mKbGxoykf9gzH2lQ4RXL6aGykBoSDx8PUJA3THQWEbX9CUKaIJcSMf5LPjqh4Y5MQpTW2O5j32a8sm5/NgGiDULSc+E8h3XJijdCZOfO1ntmlU2ZfyvzvDnOse97rT7m1HKiwsPSObCL0Z+xh9aHLLhiOnY7iTf8zIlcbhBpEn9mCMiswxjeFyGnjb5MzhEEuhhjBG/1/sZh4LjhfHHYu+NNE7vV67SKR6E7gbZFI1Ajk/tFCJ0BVVRUm058srJwB1RY7S/ZX4fFClEHJwoFlxGx+ELxeNNpQnDPeEqryNl+u3Jp8CasrFUwPDCLRuhpF3CxCTiyhpv8dNE79ANWh/6Cv2UpL+Cj2ma4gzt3bknVquoklua2MGfYm6XmvoarchtM0kGP9HyfOVYNiyZXg9RIBRCgC2HLFV3isjUiULsTSGDyJUxAHJwjk43WTEeQhUq+muqUTkQgmGhsRb30fUb97fSQOuDUmEqMjuHtCCHKJmC0FDRysaCHaoOTmmFpiJBqMrupe1yuqOYAyxhfxlzXbqbbYWDqmgZBf7uGe6d+RSwpP/Hi012uLxbEMnvQvwcZVHw0drTwbfxSxsx1+fg9aq9CLRHw59B42Db2awlYpIToF88cm8O2+SgJVEh4dqiDz2NNI9pThnfUu9aIQflDMJj2jiyBzIELGzofwABniwHQ4SeShfZEe/hJxg+/69HmLmDt2JDmeRGFwgj4K5nwKNYewj/kHB2u7qO8Q8+ERD8VNHQSqZMRo/J+OiB4mpCmsDULe+egPvW8ySxlquZRHp6axuaCBLZ6nGd03F1H1ARwRw+hSBhHz840gVZId7iEuSE2gSsYr+zpJnXMtty0rp6ZVyE6qZBK+medm4InHidDHwI6nfPJCjxvR1pcJyfYPyDxeOFbZzIBwjZCvDk0TomRrg2BPq/EZlJU1WntI/CRqWjopa7L9PpFLlBA9vKe47PsyftuDu6PLxYk6K1WWDsIClaSF6dAoTqEdkUioSZ0KRzvsehfF1heF/0sV3R2mk89+fX8G1kbBO12p730954jzxWHngr9ER+71etcDZ6hcnDsOHDhwXmwg3R5o7fZ/vneAmJhdT9KlTyJn0HNsa9KiLhaRfclG4o6+TUDtLuriLiY3/HKSu+SUtXrYr7icN74qJD54EndEeahr07Ky6lKywi7nQI2bvINWZvZv58aRcXy9pxyXx8vsjAhCtApEIhGt2hQeEj3I8KEPcbjRS7LdQGbBo/5aXUcb6vr93F4yiupWB7Mywpk+6AlSNt+GqNvESCVV8PnURfxjv475/b0MOPw4UrEXi0clFOWcNtAEs3HQu9z1XT5d3f4b1w+PRSKCBwd6SFh9NfOisjEPuLPX5+SKGsHuWv80295SC55BGRCWQcqqy5Bfvpb4YDUF9f7DN4yuetj6tKDDvugdKFyPuOAX6LQIyhEArxfl3rcJHzeI+RvkpIRqeemSNK5V7kRlzsO48zuhHR1wu13sC5jEy+uFyP+xaRFo5JKehhyFVMz1MRZEFRVCNHf8J0iciDhvea/3FdReQMaAyWDsLtuEZ8LPD6Pt10GIIpNPy3WUmTsZFKnhqaEQoz9NVx+SCn9bBSVbwdUBCWPh2PLe+wCRBhXXDY8FYsEcCseXIT/+gy/9M+oJgrrqWJm0C4W1GvPgORy3hlPT6rMm6Ohy8/H2Et7qOoRMoe01yBpXJyaJ/+cPoOuogiN7BdWJWCJ02xrjeu1na2/tJRUHUMrOQeonFkPWPCEt1d79eBiWCYkTzri7y+1h6f4qnlrhMx17eEoqN4+OP/v5Go4LfQQ9B3LAirvgti1CWuyvRvVB+OFmQd6r1AuWx2kzQXL2GaWn43xx2LnggmkIOl+IMqiYnRlOwbp2gsQ2cDnYl/Uq162y4fEKkYl6n5V3rnqErV1lJASaeO/n4p6Bu6YABXdPSOL19SdYsDqfZy9Op6DBzvFTbKWLG61E6pW8ceUASptsbClo5PHuyPXhqSmM6xeL2dbF+P5yssXHEBdael1ne1sbeXVWGtsdvLe5iBtmNvWQOAAuB0lHXuPLpJEoNr0JHhfmia+hDE2mfu5ygjY+RF3cxTyyxd5D4gBf7Sln0dWJDN0wBzwuFBXbcGc9RGPWA4QceFOI9gKjaB3xGKu/8tfKD4zRM/8XCx8PvhVjyjSUKjV3T4jmwSW5Pd7mo+O09G/tXsPLd3YPSy4U9MN7emfYglz1QDQnGqxsKmohnL5cUb2wh8S9gTFYxEEcaxNxMi30/tYiXrwklbamWtxuN4PUjaTvegxs9XinvURnzDia4mahb2lE17rY73za6AxcOl2PZUBykJJ+hiSUBz6nz8i/84E+F/P4/gR2VKMLv+zMHYwhqQKhfjYdRv4dmoqg/iiIpbiH34Uk4gwTcIxxkHGFMKm+vU7QaKuDkK66l4DuIm548RqUk14nQBXlN3y52OLBGRaFzIvg633qGD6lnvDgIESi5h4yjjYoyHQfg9X/gqisnoXlTLDXlXHTyGS/tNasjHASQ3970IcfTOlw8zpB2iiWCdH/KY1Qp6K0ycZzq477bXt1fQHj0kJIjwg842sAaO/t2Y61XtC9/9VEbm2EZbf6ejQ6W2DpjXD7NsGy4P8ILhgiHzp06Hk5rk4pY1p6GFKxmFpnA47Ui1iYJ/HzoLY73RypbOa6USl8d7Cuh8QB6tsctHd2EWNUU2G2U95s5+GpqSzcVkKLvYsUk5a/T0zm798d4k55Em9t9FcsfPtrJTP6h6ORS0lo2YPxwFNCQecXn9kVIhEVxmE951XJJHhaz5D+aClHEZMNgdE0DbiT/NCZfLKplLzaNianvsmcpFCcu48yvV8YUrGILQWNtDtc2BtKELf7WsarG5t5sWQ0t44ciVHqIDNcg8tu49LMMH44LKRp0iMCiDaq+eFgNUX9jSSHx7CoUM73B/J5amYaQVoliLzEyGyYFguCPFfsaKSdbcKjsa1ZkLdV7PF7D21y36Nni72LxUdtDBl5L8lb7wZdGKI5H9PQoqZ/vB72Ci3cbR0uHM0VXPvr5UK7/il67ya3hr/tj2Ok2cyV6Tejrt2DpKUMAGfapbRHjuXexYfYX+5bPN+85BkuqZkBW19G3f9y1JJWGHK1EK33XGgN1BwSSCUkRSCwzGtg8wJhdF/aDNCG02AcTri6d7MRIDQvrX9a0GsfXwXZd/uUON0wHHiby/u8x6cHfSqaK9MkaI5shpqdQuv89tcEXbcmGEY9gGn3v1g68z5yO0PR08ZATx6xe/4p5JntTcBvE3lSQizjkxLJTgzieE0rSaE6BsYaCFT9gehTH31OszxbO3pPcvJ6wXKKOdoZYYg5Q4dp8m8uGH8K1jqhAczvIj2CpcEfJPLzxWHngguGyENCQs7bsZNMOpJMOszWKDpbHqFleXOvfWwdHaQ0rCW/Nq7X3/Lr2lkwMx5n3Qnig+pZeFQwduofqWdTfgMF9W243F7cZ1AAdXa5ud5wnMBAPbr8TTB8Pt7GAkTTXhJ8L+QazFkP8NQmJSAUTds6XZgNGfTqp+t3OegiqJ30Djs64nh56bEe8l+0v5a8ehsvz8ngnU1FuDwebhoVT05VC9FS343qDozloC2YQzXtzK+BEJ2Cn/tvwnTwLW6b9QPRwcl4vF7Km+28tbEQmUSEKGIgr+fWs7O4jodGB1Pd1s6Cn4WxcRf1D+PhkS8QmfcR5iH3ExqoFZQhox8EyQjhB9HePU1owHVEKDq5Z4iBd/bZCAtU0mh1YokYD7dsFKKtgAiCJNWE6YK4IiuaJfuF0WGFNjWeqKGIy7b7Pg+RiCp5Ai6Pi4+3l3KsJoirh3yOsaOcLrGCow4T6WavH4mnRwRQ0yFl64yNJFp2Ed6wldaEi8lrDydJ3UlYoFKI0pbfBSWbfOea9RaMfQQGXIXF0ky9OAy9MQSvy9+DHHOpsHi1VUPkILjya0HvHpIGml6qXPB6mZms5Id8Ox1ON7cMMzFDc8TnXb/jDbj4PaHRy2GFrS8j72xhcNUuBs/9XHBNPLFaIHGlHgLOPtzAZDIRZC1ious4E61bwaEG040Q8D+f5PNbiNCrCNEp/AIjrUJKtEHoMThc2cLq3FqsDhcXZUYwONaAQiaBkD5wyUJY/YDwpBYYDZd+eMYOU5fbQ0O7A6VM3MsA7Zyg1Av6e/tpnKA9w3f1OzifHPZ7uGCIfPXq1eddumPUKvj0oJLLBkaSU5Xv97eJQWbYvIBZw1ews9h/+vmVA4IZ2b4OSc0a2LGJp+Onc9D0MAu2FBOglKKWCxG+SiZBIRXjcPmikNsHqonedY+w8l/xDWx8GpEpHY7lClFGl516r56qNl/eUySCIlkqxklvEbzrWaHjre8lIBLhlStxS5VY7E6/HwjAoco2TjS091j2vrWxkBcv649EK6Oz75VI5XK2Gefy+lrfue4bqiGkQDDoitv3PIq4V3hpq+/R9qlZfUmPMnCgso1ZaVp0YgdvbPFp31fk1pEwfiyhGcPJ0iUTGqyFSz+CtY9BwgS8ly5EVL5TiKSLN2I8/DXzBt+LYeY81uQ1YFDLCAgIgHAfAZ28F56dFsOTg124RRIkQYmI7S/Diruh5gBEDcU1bD4Zxz/hl2gpbZPm8rfNnRTZjby1UQZ4eO2KUNodXWgVUuYMjiLaoMKgkfPqmgJq2zoxqEN4csZTOKtryDYeocOp5UR7IkmOo4hLNgkdkSGpINfRpTQiXnY7Hq8XR78byemU8faPR3hofCQXh3kRi0VCPeDbq4X00knMfgeu6W6xb8zvnSrJvJJBO25n7eC5uORaFAkmgg1jQf2mkLtNGAuaENj+ur+nSnAyFK6Fo0uFJ4XIwdB/7tkLdU47hvKfYetzQgohearwpPHddTBv1V8e8UboVSy8bjAPL82huNFGlEHFK3MziA3WcLjSwpUL9/RE7Iv3VfLlTUMZk9JtjZxxBcetOvrEmYTF6QwdppVmO//ZWcq3v1YQqlPwz1npjEkJRi79A639+mihrrPkhp7pTox+UBjU/Qfx3+Cw38IFQ+T/LRQ22qhp6eD+ScmsPVaPVublrgwYmPMEtFaRpazlqiHRLDtYLbSxzw0irfAFJEXrhJt+6gtoNj9Pn6hRpEeMJDFUi8Xu5P5JyUhEXj66NpNvfq2iqtXB9aleJrV+B52t5I/9kI2lIZSGLGRKUgAjjj+P7rgg5erTeILXLl3J8aYu7E43iSFa6jpc5OgnMzHjhFC1L1oPR75HFJRERMp0+kbe3Ou9nT4BCOCHg1WsEIvpY7yZu9Q7MEYkMiujGbPVyd+yYyiprueD2DcYkdlK/yMvcsOAAIYlhlDXYifCoCY1wohKIWNCHxO6VgufHnb0OsfKo428ddUAkk06YUPSBDD9hMPajKTuMNJT7YGBoNyPUQ+bQU1LB89e3I/G9k66PB7sTjfxQRrBs8JcgnL1QyiLNwqLwOAbBcK6+lvctkZoqUS6+KqeYxqOLeWbud/yaZ1QDL1zbCJjk0Nosnby4JQU3t9cTKPVgUYu4c5xSSw9UEmF2U6Sp4QBR+YLJCwS0zTwbuyJo9BmXiWkV44twzv1RWRLrxc+YyCseB0DpnzFk7OyOF7bzsEKCwOi9Uhrc/1JHISp9EkThck5Yf1g3mrI+RavpYL6xDm4FAaCjMdR2asoDLyIHaVq/p4YB8YbIetG4RguJ1z2Cay+X2hkMiYIRmIbnha6JQ9+Ifw9YdzZb/7awyjXnNLecWKNoGqRyIT0wqlEbqmA2kNYJXrM6ngCDaEEas9gsPY7GBRrYMntI2i2OjFoZITohGNsyGvolXZZuLWE4QlGgYhFIrYfq6LPyBlnPK7b4+XL3WV81u1xX2Hu4Nav9vPDHdkMOoOvzlmRMg1u3y50q2pDhcVbofuD7/R/FxcMkcfE/HembU9ND+PGz/exv9zCmrkqQg+/j3LThp7VOGnrPaQP/gbj6HgygyF1z3wktd0NG+W7hIp61s0YqzcwKW0SD/1UhP0Ue9h3Jyp4T7MEZ1w/1LteAUcbxaNf55rtQZhtgvZ36aFaXppxH1ea4kCiAK+bUd5D7LElsfpoI1IxDI4zMkzTCXve838DzUU4g9KQir2MTw1hc4EvOr5mWDQbjzf47R6oEuxj/7O/jZFXzGBEhJGZXTKsHS7u+fZwjxJEKhazaN53DCv/iUFrTrEdnfgMDJ9PikmHoxMSA4T9ZRIROqUMs81J/wgtqSb/hpJcswhFTSnJ3t6FXWRqBscFM7a9g0d/yKWjy81toxPYVtiI3eHihVmT4PDXUNytYvV6YP+n1EVPZ32RjKNNsMD5pX9LvdeDuvAn5o55ibmDoggPVCEWi3C5PXy4tbjHksDmdPPGhhPcOyGJE5V19Ml98RRljYfgg29jiR4qRLw5iyEsA2/5Tk5fIqNLv2d/UibvbCri/S3FvHP1QGZwmlcLCPWCU0fARQ4S/rXX01pUyNoKL9vb52O2dVF8wMoH152hCCiVQ/85QiHT3gwFv8CGZ3xqGICSLZBxee/Xnor6vN7bCtcLRdlTFRqWMvjuBo72f5RnDnvZX3mEPmFaFlzan8GxvYd+/x6CtAqCtP5pjzN51jjdbr+0+Nk4obHdweJfK/22eb1QUN/+x4lcLAFTX+Hfn8B/i8POhAtGYT9lypT/ynmGxBl47fJMpqWHYe3sQlmyRvgi9TEgkSESi5gmPUicuAG9oxZp7QH/A3RYQCrHGTGcaQdv5/tpHvRq348gQOJEevQ71DKRkBIBjpHkN/0G4NXtjTR4DbDlBdj6MqrV87kntZV7JiRx2eAoGlo7MUh7R7/e0L5s6Ejlqi+PkxGl54kZaVw/PJaHpqQyMNpA6Sk+4DKJiOEJQRytFq6jwOzBWbYXcUsZ+fXtfv7kLo+XT3ZX4zq20u989vz17C2pZ8m+SnbboxgVp2XZ9fEsmB7HlYPCeGZ2GreMikfW7arY3tHF/jIzsupfSd0wDzGCKuZUtIx8kpuXVbFobyV2p/Dj/Xh7CfPHJXLZ4GjK2kU4m8v9XtNlGsCHhYE8tbaKEw3tZ7QMtblEPPj9ETblN2C2OSlptFJuEWaWngq3x0uXx8uIcBGK6t29jmO11OGWdxOq141X3NtLxCuRc7Ir2u3x8tgPudgNKSA7TYs98AbBI/wkPB44sRbRB8NJXTGTe4tv458ZrTS1dzIvO46ss5GQIVbQbHdYfCPlTiI847dfdxJnyvsa4ugwZZEvSiK/to0OpxuqD9IYMY75O5XsrxTup+N1Vm78fB8Vzbbex/gfYHJfE6c/QN46OkHIkXfjbJygkomFesZpCFT+McngX4n/FoedCRdMRP7JJ5/8V/JLWqWMOYOjCNcreX5tM8/MXsYxi5icJgmDMmBImISwH+dwhUiCe+IzQqRy2lBdj0KPWZdOxPZ/09f9PK9d9AW5jR4iA5WkKQsF+9lDX+OYuwjv4UU4pRrAX/vrdHnwyLW+nGlXB5q9b6BMfQmVTML8kWG8fKiap0f9E8OufwtPDFoTHeOe4YFFQvPSloIGgrUK9pWbae90EaiSMX9cEl1uN26PF41CyifbfTKzGL2cBrsXkcpIQ0VvG9VmWxcenc+VzxsQzfLYJ3niM8Esa3i0mpeHd7Jgp4KDp9gFPD49jb6RRjpdbj7cWsywaCVJjjJ+mbiO7wsc3DNqEH079iNvLcMaP5X9riQqzAX+n6kX8uusvLe5iBijmkGz7yXt2BLB3TAokfqwyXz9i/C+c6paKZl5LX2K1/gOIBJTEz2bbINQENuQX88HW4qZlh6GXi2jxe7/HcolYvLNEtxRw5BU+pO5WRJCsFYmzOqsP4Y462bIW+bLUYvElMXOpabFR6ZtnS7KxTH0uX6FMGi7+YSQChp0va9bsbNNKP7u/dBXzGyrof+ue9l021q0obEoTsvv1rZ0UNPaSaBKRnywBolYLHiSH1/h87gJ6SOkb34PEYNwhQ/2BScSOdUT3+WlX52s/F5QF106MIKHM8KpUzuoMPsvFm0dLsrNdmKCejs3/lFkRuv55tbhfLGrjPbOLuZlxzMiMchvn7NxQqBazj9m9uXGz3/tUaD1CdPRP/osssbzjN+63pOF3RWHa2jt6OKyQZEMiTOglP119HvBEPl5h90CzYV0uT3UEoyEAIIMBp7L6erxC/8MuHRAGPdPeB9D+S+YvcGYRj6Mctu/ew7jTZ2B05iCzCNm5+yt1HTpKKnt4rNdpaSZdHiGJBE+axvZ8iJaNPH8iwcZJgtCJWvyGzLwtxExvFxtIHvoYsa0rSb0yEJk5kKGBLv452Yzo5U2lh/roqBxIC9MW0aCzIJXouCIJxGnWyDB3OpWnpzRh00FDXi9gpzvjQ0n+MdMoVDz8baSnqk8M/uHIZKr2dGegNorYUIfLcsO+UscLxsYgXyfzzOmot/tLNjh05bf1qeLslaPH4kDvLHhBNP6hdHa0UW8pJ4MhYqfFNN5aqVQUN5UAEZNX765JJs0rYNkjYlgbSkahZSp6WF4vF425NUjkwghWoXZznGrjrTprwie2nVHCNNG8ML4aTy8sR2PFx49oOGFSV+RVvU9LpGMmsSr2GmL440NJ/B6QSoWLIVXHKrm8elpLNpVzF393Bi8FgJNcTTIdSQOjsLT/m8k310l6JRFIsyZdyCNGoBS3iRMatrzPqLd79I153O8JVtxuty0Jl5Mrj2GT3f4pKZRBhXB7gY8IWm0XfIFKm8nioAQ3+Diyl/xrHkccV2O4C0/+TnY+C9hgbY2EORuAqm/6+ChCgu3f3WAhnYHcomYp2b1YW5WFKqwfnDzemjIF4IGU99z01froziccj9Z4xWCGiQkjU0lalbm+hp2fjxUw5CoZIZJi5FJRH49CcAfkyl2dQoj5M7QWi+TiBmeEMTQOCMer9dvbuq5YmRSED/OH0lhQztahZR+kYFEGXqPHfzfRm5VK1d9tKdHBPHjoWo+/VsWE/v8dV2gFwyRy+VnscL8s6g/Bj8/DOU7kelj6BzyHM/uM3Db2CT+/t1hv11/PFzH9P7Z2FKG0tjuIDagL8PnDkBmKUKhUCGp3ou5cC+PVmWzvVwgwmCtnAenpPL86uPoNXKuSpMhPvYlpoaj3DD9F3IaPbx2RSarcmuotnQwNjWUkkYrq3IbWAbcOvgiHgndjSV6Ep3N5bwy3oQnfyumgKEcb+jgkmUAghLjnavVROpVVLd04PXCt79W8PTsdI5UtaCQiLgkRUaL1cyru63cmB1HlgncUjXHGx28v7mEY7VCmmVInIH3rx3Ie5uLcbo8zMwIJ0XTATNeFkZzNRfSFZjIW+OlRInb0AeZCOxqZEezDvBvY3e4PDjbmxGLFExXnyCnYxyf7fLPX5ptTvY2q0j76WJiZ7zKB9dNZVVOLV/tLkciFnHLqHhqWnwLRKLaCmte67EfkDYXcUlCPltT72NVQTu5tZ3cuVPLslvfxtNcTGubkwVrCntyrC6Pl3c2FXHP+CTkdPHV4EICNz4qRNUSOcpxb1GrmMZNK9qZN+ALJppsiBVarDIj4ToVIuMgoWM2KgtEYmSGBOh3MXJA5fFiPF6PSi7B4fIQbVDw9mg3ncfX8pJ9OMuOthAfouHhKVKy4gyIzKWwaC7ik46LJVuEhaPfZZC7BOQa3CqjX87fbHPw0Pc5NHQrk5xuD0+tOEb/KD0DovWCXbEh7g//FI6WN5M1fl7P/39ZuafXPuvyzVwxtR8P2Tt4YZtPwXXLqHgSQ86hcaitVvCkOfy1YJc79FahMeoMEItFiHtVIMDpcmMKO/t4OalETGa0nswz2FX/b+C3OGxbYaOfkg3ggy3FjEwMRin/a4ZnXDBEPm/evPNz4I4WWHGPIFkDaKkgZdMt3J29hPymM+f7dC4zkS05tHdGcvc6K/dOSGZ2RBBxSyaDWMK+Ed+zvdyXmmiyOtld3MzAaD27ipp5KlOBpHgDOZO+5ebFJ3py0aOTgrhvUjJ3f3OoZ5tIBC6Fnl+GfEa700uqxExc2wnkeg2fTtdw5Y8ubE43YhHcODKeL3aVcs/4RNbl1bOvzEJEoAqv10tWqIcpHWsIOnYUooczZvZAJMWfI1u3CnvsRDbJ5vWQOMC+MgvZicEkBGuQSkSEy+wEKWQ8dyyYeZf/TJe1mVjbUZK23yd0M65+BLJuJlkSRYAyws+xcXqakagt9yHPmAv6cHZXuXoC0VPhEUmFitTaJygfN5wvdvvy4G9uLOT+ySk9UWC0u9LPQwZAVrKRp697gpHpsWjkUvpGBBBsL4DF0zg+/FM8Xv/b2e50ExqooJ+0msBfHvOlRtxOYnc8gmV6Gvm1Vh7rlrk/MELJvSWX0mFIwz7rddSmPsLsy1PhtCGxm5maZKTfbQOw5G0irH03AfmHeUr+KN8dFaSbjVYH1326l29uGYap4ShRp9nm0nBcsB4WianI/jd7S2XMDnT1PG43WZ0UN/a+P6stdoHI/4fw+53ZmhgRr2dXsb+GerjJhVQVyLUjoxmY7KCqzY0pUEV6RIC/T8qZ4HHD3oWw8w3h/zWHIP8nuHmD0Fj1OzDbHGwpaOTrvRIqiqkAACAASURBVOVE6iOJLrcwMEbfM1npQsZvcZjb07u/xOXx4OHPOc+eigum2Ll8eW+fjL8EbdU+Ej8JdxdR3hoa2jtJDPHP92WGq0kr+oT0bXdwXf58Ppqu451NRbgddoEIlHqK2nrfzCfq24kL1jAkzkBo7Vbcpky+KNX5FRS3FzVTaenw23bdsFh2l5i594cCnvzpBJcvbWKrKx22vUKyvJkf56Ww5KZMFl+TwJzYTtL0XtLkDbyRXsTKK428lpLH+CiIlzZzRJZBpTEbW5cH8Z63ke15G8wlyFuK2V3aexrSsZpWJvUJ4c6hRq48civRrjKuSJOxZG8ZG/ObkK17TPDu2PKiUGA7+AXxGhdfTZcxJVlHpF7FHdnhPBqZi6psI5K1jyJSaqhrbuHiAZF+51LJJBh1KpzBfXGH9mPJobpe15Nb1cLw+CAempJKVbu7198RSyg1O/nXT3l8uaeczi63EPm5u4gUW/zG+wEEaeRIxWJ0rubeniWOdjpaBdLVyCXIJWIWHeukOf4iVGUboHJf7/PX5sLi6+CdQfDdtUR2FtJv2+0EH3qH2phZLM3ztzhwuDzsLmmmzHqGqEsipzlkCOtGL+Xm/dE8/MNRCup8dRSDWk60sbeJVVigb5vX66WxXeg8PlcsX75cUNIcWgQfZDPT1EJysK9omBaiZIoyH+py0AYGMTQ5gssGRzMyKRi9+hyemlurYO9p1gydrULj0jngp5xaHliSw8HyFn7KqeXqj/eQd5rJ14WK3+KwsSkhfuMSAe4Ym4ha/v9hjvykefpfDoVO6N7qbPHbbJMEsjKnhkXX9+XHnHp2lNkYH6/hakMehm3/AaAjpD8hOhUPT42gQS4mJjQDWeMRMg291SQjE4O4KbaRGMtu5FoTosTHuK+hjNsjRRx0RvPMDjsOlweZWIQpQNGjpAg+bR6o1wvPbGwgdub37GwOYltxM5sLhM8mK0rDyyNcJBR+hDthAoH7/01pv/k8s76OrUVOQIRRk8qGq/VotvhGp0objzEpxc0h/2wHWTEBDDSvAZsW77h/ID26hNTCX0gwpmIb+wzsbxKKvd3qG2yNsOEZMvvPZcHoa2lFQ+K6mxA3dTdXOdoQeTxc2T+AN/Y08cjUVPaXWzCoZVyWEYzCXktL5ARCq9eTatLya5m/NDHVpCMjMoBKSyeHO8NJjhyB8hRVSWvmbbyw10Fnl4fSRhtPLDvCj8lOJEDC4Vd4d8qHPLS5g9aOLkJ0Cu4cm4hW4kYv7hJyte5TlENqI2K1kUPXWlA35mKX6jki6YuiSohOpWZ/qwXMpYJda/QQCIzAVZ1DfYsd+egFhOx7BbmzFZ1S2quo6vZ4WVKpoV/qFegLlvRsr896mGvXySlq9D3ZVbd09KQJQnQKXpmTyS1f7sfqEJ5w7p+UQmqYoG+uaeng+/2VLNpbQXigkoenpjI8Ieh3c80NDQ1QsRdWzAcgoXY1i/qoKVSmIwKSrFsJ3fkOXP75WY/zmxCJBQ8WTvuNiH+fairNdj7YUuy3zeHykFPVQnrk/14R81zxWxyWEaXn29uG8/WeclrtXVw3PJbhpxV2/ywuGCI/b9DHwIxXBGOcbrSkXsGSCh3vT5Yw4Mfx9FeHUD3rfaJy3kK8VVhVy7MX8HxlJuu+rUcqbmDe8GhUYxcSf+QNBpl/4a4RV7FwbyMuj5fsBAO39HWTvKR7gseUBfDj7cR0E2CyNpyA8R9w13o7bZ1dXD88luqWDsqa7X6Td06iob2T/bY4mju7ekgcYH+VjRUVAdyXOAGvNpzq1BvIdYSztciXojDbnJxo7mK4WOrrVLPWM0t5hB3xA9ldKpDnlJRAZjt/IfLwmxy8aAO11mbUSbcREjoS0+F30Z9YKrTEy7UgVfrkbk4r5CymIOJmVHIRzTHTCGmvBkc73pBUHAoDg9oO8tCYkewvb+HiRAl95TUkr7wUj1xH5bTPacu8mKmdJlbk1vaYRYVoFajkEu5YdAgQXBs3pD7DtMHFSOtz6AwdyGsF4eikHpZP7yK8dR8dKhPuxBlIfl2IuL2KyXvnsWrY3VSETWFDJdTXVTPFKCO3JYC+M95Fu+5BIRpVG3FdvJAsewmSZXeC14sCGKMNFRqPjoPo1JyupUJoFy8WWvarxr3Bfzw3smhpIwZ1X/455gcmFL3Ik6PG8PA6X065T5iORquTlQU2Avtfw+XjpqJ3N6MxJfFarpKiRv/gIvI0G9nhiUGsumcUlRY7BrWcpFANSpkUj8fLN3vLeXdzcff94uBvn+1j2Z3Zv5sv1ul0wgDnk8hbQWjaDEK3X+/bpg0FU7+zHuc3oY+GcY/Duid92wKjz8m3pMnaiVzaeyH6v5BWORskYhFD4owMiTPi8XR3Af/FuGCI/GxjjP40+lwMV2nw1h2hIzCZUncYs9rtZIuOga2J9rjpPL7FxgvxQ4hhOQRGscw2kHWFQjrC5fHyya4K0sL7UpTwOGGyDmYFBxMfEUyyp5REbRvald1dhqZ+ULXfF8UCYmstg537eHLGXHYUNbH1RCMRgUrC9SpCtQrEIvxMvC4bEM7WQjOiM3zhGyrh9r4xFHmjMKjrqCgx99rnvVwPQ4bejuSUZiJT6XKuyJjAA2MjkDcXkFD6EbqCXHbO3Mi9S0totjmRikXcPGo0sQOHcmlgIYrQvngtJXDR24j2fwYVu0Gq4PClW3htSzOHK1sJDxzDC6MmMqbsbcRJ4zlY5yTL6yTNtp+Bhx4RFhOH8MQhtpux1BRRGjmO+xYfYl52HCJEiESCGmLXKfNTv/m1glHXDEQaFgJ73qVansbaUhdLR5QRvfnBnv28eQm4blgJu99H3GUjJDyWBns9s7QthATYeGJXEtuKOgjVBfHA4EVMjpMQFBaL1FIOWxb4GzNZG0AkwjX6YWQxpxgglWzuIXGPIYFFLen8Z6+Qlqlr62T+qk6+uv51Bmm8fHlTIjmVrbi9Xkw6BYv2CuPCvjpi56sjKi4fPJwnhqVxmdrK1rJDGNRy6to6uWdCEimm3oXEuGANcacNam5od/DlHn+dvdvjJb+27XeJfOrUqZDf/TlLFeBsh5ZKYfJRxR6hONn3IghKPOtxzooB1wrujwXdgy1Spp5TYba2tZO5g6N4fb3P9TNAJSUt7P9Gl+W5cNj5IHG4gIi8qKiIoKC/9nGjBzIFaEIQbXkBNTAQQCLDO+YRAMqNI9l5oI3vgvpzxdTPaFDFs3pT77zc9uIWDpRbyI43MNxl56Ef8vhgton+nnKfq53W5JsYfgqM9lKOtbSy9YTQiVnT2klNayfDE4y8f80AXll7gppWB3Mzg7lmkJFblhRz8QB/SVmAUsqMZA1tCjViu4tgRyXpAb2bPBIMCiyR42gYm0Go5RBtAUls7kjmuRUlfHGxHq3OyOH423D1C+PZXypo7m5Wcnm8LNxWwurrIlCu/Seik57T+/+DZ/a7tA28k0Z1Ek+saSSvViDn2tZObl3jYNW850mtW015h4TXjsXz8SQJKkdb79y018uJ+naeHCEnRXwQp1jJ2kYjbx60cs+EJHKrW+lrUpHf0IlOKYOGPGjII9bzEe9Pf4foDf/2O5zIUkJRaSmPNd+OQipGusvJaxFbCDv0Jmsu2s+2IoEUGtodPLbFwX9MWr67zYShZbtPy30qJAqko+7xab9B6OjtRlP/W1iys/e9cbTByWarkyNVNcyfkMQtX+zH7fFy6+gEspOCKWuyMamPiTEpwRi0CmLdHu6flMKOoiYuHxzFhNRQVHIpXq/3dyNQuVRMkEbuZ30LoBF1QtkOgYx/w5GxsLCQYUmTwGGl3jCYfIcRJ3KSwvWEDbqd+nYHCpmYs+tFfgdqg1DITftjg4gj9Wo+31XOo9NSOVrdhkEtIy0sgKRzUcqcAzq63Cgk4vNGpueVw34HFwyR5+TkMGzYsPN3grD+cMkHsOYxofgSkYUodiRIZMhFHkQiqHIF8NgROS0d7aSYtBQ1+DfxpIXpWHesniUHa8iM0fP4aAMTVEWId32Ap//liHO+her9kHWTYHh0CipDx6Fu7F30kkslvPDLCT6dLEJdf4zQkh+Rlk3jkfFzOVDnYmi8kaPVrdwxNhGbw4XF7eWILYAktR2pLIqsuvX8fdg03t/fRpfby4g4LTdkarBV5jBncyRhgdMYnRRCgFrGjSNdSGihz8FncYx6iBJbE6XN/o1BarmEyI4CH4l3Q7z1BfRJk6jV2cir9c9Xdrm9lJeXkLrnRS4f7WHUhGxqbU60ExcgX/94z35uXRQHO8O5KKKOsBU39AxAzogezdx5r6CSOLlTtA912TrMGWNRaiLBJXxm0qbj9Hfm9ET3p0LmcXK4ykeue4behCTiWpo7pIhFMDo5hEi9il/LzJQ22ahv68SgNgiR5+5TLBBEIogbiVskQXwqocaPhlzB51ylNxEVKOtZ/E7CqPCwaG8dL8+KJY5y+oYoOFLfycfbSwhQSpk7OIqLBoSjlEnpcLp5bW0BSw8K8tVVubVUt9jJjDawZH8l0QY1Vw2N+U11ilEj54kZfbj1S18RP9agoH/zWsjvNvuKHipEw6chJyeHYSnTqBCFc9cWCUfqhCeLAFUtz8wW88CSHIwaOc9enM6UvqY/ZkD1J5EWpuOqIdE8s/IYITolIVoZ1wyPRfdHtOtnQIXZzk85Nfx8pJah8UauGhLTU2v4K3HeOewsuGCI/LxDpsKRfiXN+sHUNjZTYA+gD+EMuHox8QWbuH7ALIKCtaw43D31ekAk+8ssPTre9IgAtAopySYNwxOC6XB6aHYp2eXpyyB1DB2JlyOWRhBa8gNOqQZR9n3I9i0EiZzaQQ/wdlEokeFyRiQGsbu4GYlYxGWDIimoa6PcbCe3WctlexcIypiWEqZNSSBar8USlYRifCJ3Lz7cU0j7ZAc8e3E6IYoIRpr6ck/ZZ1w0PhuHVIM4OIyjNVamBwdz/8ggZLoQPt5WQr9gMXelWUkSm3EPugHFvo8IUcUSbZhIpUXIfz80TMNMbT4Bjt7FXJxWkKnQOuoJUAX1igYNEie4HEg3P0f0hKdg03OQNhvXld8g2vMBZuMA9usmsqcU/mZ+DYyJwhSWrg4UMhWRokZkW/7dU9w0Vh+E6s0w9gkwxIOlFPmBjwWHv8Pf+E4sVVIsiuGkBTBAp72dXYWNzBiSxlOz+rIqt5a1x+oYnRzCvOw4IQ8bMVAoYI64G46vBJWBpolvsrUhhMWr95AUquWaoTH0j9ILDoR9L4G85ejaCnlkcAh/qxPh6s6HpYYoSdJ0UtXSyTvbq/gk4FMWDxrKzcfS2VPRQVpYANcNj6Ha0smJ+nZcHi8BKhkSsQi3x4tBLUMll/H3xSd7GppZcbiGH+4cQd/fGMAwJjmEpXeMILfCjMFVz0BXLrGKDuEJaMVdwqzRKc8LWnW5LzWjVCpxt9aQK0rjSJ1PdtjW4WLt0VoGxeg5WNHCPd8eYvn8kWdP1dgtgiKsPk9InURl/anBDwqZhMsGRTEkzkh7Zxelxw7S91yHQv8GbA4XC1blsS5PWLCO1bSx/lgdS85lRun/IVwwRJ6dnX3ez7Gv3Mx1n5z05rZwwwAzqQFbkfeZxj31ZWzw+vyE31h/gutHCHrlSIOKfaVmnvkpj1fnZvDu5mKKG4Vo/SPgqfH30ldi5IVSOa+MGkuArRS7MgrHzB8xaFXcvLpNSEUUFDM+NZRX5mZgtjlZmVPTMz+xwyUWPF/cQsPKXncSRkk9EzfOZvGAz3upIb7eU05SqBZHTBozMk04LS5yLEpeWtaMtdNF6pQOJqcYueKbYkxaCQvCdxK68ZWe1zsH3ohqwFxe7RfJrYsOMzJayTzrR2hzVsGEp4T86SmmTI6s25HaGojWO3luShT3rSzrSS9f019LStUnvouzNYJcC/k/IUmaSMnMb1l+uIbiGiuPTTAg2WiF+DG0V+dTFToWpaODWGcj4tN9T2oOQ+0hIcJ0OwUpad9LhEXg0Nd4DPGYs+4nryaMiUn1XJXQicbTTrKxkUEpZsrE8PKagp6O2uWHq7E7XcxOEENTGaROFxaofnNAF8EPh2y88Iswj3JfmYWVh2tYdkMyqSYtXPQujLoPEDNi2yssn3oJJ7pC0IhdpLuPs7VGKOb9WmnHMn4oUZv/wRfz1iJxepBUb+Zoo45rl1T36O9jg9S8eeUAypptuD0eluyr8nvrHV1ucqtaf5PIFTIJWXFGsrRmeP8iwcogMBJKtwo7dLbCyruFXHes8NuqNNuxRg7j8p+bGRgdwN8nGnlnU2FPfSa/3kpGZCAHK1rweqGkyfbbRO52wb6PYfPzvm2p0+Hi94VBGucIt8dLSaOVSoudYK2C5FAt0UahO9Nj/vP+3hVmew+Jn0RVSydFDda/nMj/Gxz2W7hgiFynO78FDa/Xy5e7/AtEUyM6We+ayrs/dvHNtAgyvHakYiHScrg8fLK9lOzEIA6Ui9hR1IRBLcPmdPeQ+Em8trOZZ2abeGNEB4k/X9sjc3MFpeAZfhdvDFCxPCqKoxYJs2KcBEqdbK9p6yHCackaZkfZQXE/WBupUSXx0IY2XG4Fyya/Soeld07P7nSjkkl499cWxk9ox1S1nb1ts3oIv65Lg8LaRaPVwT+HKwnd+x4MvF7wdXa7kB/9gfY+VzFMV8+qsbWoQhPQLlklHPzXj2DSM0Kxqr2W9v43UGIYRUrrDlTbX2Sa/mdWzH2eckcgwaIW+pR8RmDxKWZbWlOPYkZUvInErBt5cEr31Bq3C7JupKjVwz8sE9iTY0UhFbNujpwzO2l7YfuroAgQBgu4XXgn/IMCwzjePuDgl69amJXq5e2YHWi2vyIUL5V6NJd8yh6Lw88WAWD98XqaInMJ3PEgHWmX0jn6CQyR/alt7eC9zTl++9qcbo6XVZG67QW47CMhigccY/9B6p736Fe4mo7QgRxJvot/rxW6UrOiNOgbBA26wlwIq+7DZcrkP5Wj/ZqoypvtHK1uZdHeCq4ZFkOfiABqWv29Tc4pl6szQdRQiBwI+z/r/femExCbTXtnF//66Rgbut0xD1ZYSAzRcOWQaL7tdhEcmxLCpnyfSipII+jGSxutHKxoobWji8yoQPpHBSK3lMC2l/3PVfCL4Lkee+6Etim/gbsWHeyxtH1oSgo3joxHo5D+JZwgEYt6iQkApJK/Pk9+vjnsbLhgGoLWrl37+zv9xZDpI3l2ZweFjXbGftfJ5vJOXp/Tl2iDCpEIxqWGkJ0YzI5uNcXDkxPRiXuPqerscqPAQWzuG35aZWnzCeTtVaRuu5u7Ulp4YqiETGkFW4vMHKq0kBCs4T/X9ecdwxICll4BW1/GW7oVsz6Tq9KkLBlZQ3jrQVISYns1FMzKCGdLQSNewKqNp8E4kHlpAolLxCJMYZGESDtQySRoRZ0w8Slhpua2V4VIauC1BChEsPoBYrb8nZD6Hb6DW+th7RPgcVEx9nXGbU9DYilBtfEf4LSiaDhMxk8zmW35nBGKEvSVG32vjRsjOACOvFdQL2RcCdWHBL/r1iqo3k9nQAwfVkSzp0JYEB0uD6/myHCl+HtPe+PGCH4iom4nSXMJRAygrLyEy75v4ud8IXK8JrYdza6XfQqUzhaU6x8jWNU7v6uRS5HjYt/kpdzWMo+Lv67kg80nsDlcZ/RzF4u8glqnaj8g6Lev/bGZOZVz+HHEUr6JeZYrVwlNXqE6Bf8YYEdb9JNwzR43eFw4AuLIa+7dxVff5sCgkfHRthKuGOSfktAqpGRGnYN2WqGDma8CYj+liScwjuJRr7NXPoLyZhvlzfYeEj+J4kYbpgAlIhHMzghDJZNQZREWpOn9wkiPCKCsycb1n+7lwe9zeHZVHnMX7mZ7YZPwtOY+QyOS89zdEastHTyyNMfPl/zVdScobBDqIH8FJ8QGqbl+uH+IMCBaT3LoX0+6/xscdhIXTER+viESibghO9bvMauyU5i0A0KE++r2JtRyCx9cnUGwqJUaj4Hjte08PDmJtGApI369l7qEOegUkbQ7fNHVRZnhtLS0Im0t63Vep0jOrov24BIp6WdZx62H4zhSK/ygKs0dHKluYUmfMEzdMwpF5hLSyr4kXhWKZvu75I37mKdXF/PEjD5sPF5Ps83JxD4mShqtNNucPDExigc3NLC/0shHVyYyNqWJi9K0JO55AFl7Nc+PX4hbJcd7cAGikwNmnTbY8iLS6GFQ1d3B2FIB0cOhstt7w+vF3dHKtydAIhJhMu/v/aGWbBacG7NuAl24EIXXHBAKygADrhPIuyFP2C+0LzQX0WoYzLoSwYFwZpqWUBUsP+Egb+ZDpEdm4a07gid2JPKYIYIKIzhZSDtJVZA0mbr8cuxOHzEa3Y29Lk3cXEiAyN6T8z2Jh8aFUxSUQE6Dl/QYN7/uLOOltYW4vfDA5BS/ie8GtYx0bzEtiReR60whb0sxBo2MFJOOxfsqub/GxqAYPU/P6kuc1kVqw8+Eb34RJDI8kxcgrhDULpqmXK7ODOTFzTY6u3yklRiqYWWOUPAUW0r4ZJqa5eUyorQwa0AMqWFCftjpclPcaKOx3UGEXklCsNY/Wjelw5iHIXE8LL6GLl00q/q/yWOb2nG4itAqynjvmoG9vz9gYFQA6+8fS0SgkrJmGxlRgQSqZfQJCyBIq2B7YRNVp7g8er3w0pp8sm4aSGD0CDjVOVJtFKZenSNa7E4s9t6LwemTr/4MFFIJd01IIivOwI6iJjIi9YxKDiZE9z8YC3cB44Ih8oSEhN/f6U9iSKyRb24dxuK9FUglYhLDjFyaGc4Ph2t79ulye+h0Q6NLzKubC7hycCTOLg+hjbtRV+8gofkIX09ZyIdFBo43OLgsRcboPgbuX1HKjNQrCDn4pt85j4rTmPdtPrFBal64eDpHTvM3Lzd3UBIxC5NGBDIN5K1AWrEDacI4kGvZaw+jqLGdBavzmNwnlKuHJrGzqAlHl5vX5/Rhe3ELB6vsvDNJTf/KRXymOAquYYj7XwprHmXWsQfoGP8coobeQwVam+rQRA5BljRemAeZORRiR0DpNizho7EnzeTDLxpRSMXYtDGcnrH0BqciihsjRGdqgzAuyw8eOPajb3EAmPpvdJ01TE1N55boGhJyX0TW2MiNA26gnTFIdr6Oe8ityLNuEHw6Nv7L95SjDYWUaYQYjSiklh4joiaJqde44a7gdNbmNXDL6CQ8Xi+1LR0kBilYmVvHM2uFVEKwVs79k1J4cU0+724uZs3fx/DR9YNZfbicBL2EqbFi4uwVfOh6kFeW+NJyQ+IMXJQZwcqcGg5WtHCi3sq3tw6nXTkZSUQmGn0oGqkYlAE4RUpyE2/FWuPm7nEJuLwivtxdxqUDozhQbsHjFRa0uMhwkn99ikmRaYJsL1yI0B1dbpYcqOTpFcfweIXpR29fNZCp/U4by6YKhMQJcNsWSlrEPPx5cU8x1upw8fXe8p5rPokB0Xr6RRl6hj70jQjslZO3OvyL2gDNVicOkRwufltQ/eSvgsgsGPeYoB8/R4QGKIgyqHqeAgDEInocDP8qTgjVKZmdGcnszMjf3/lP4L/BYb+FC4bIx4wZc97PoZBJyE4MJjvRN8TV0HYcndjAsrx2Yo1qLhscRWGDlYsibawaV4uk6wS1XhmKoGihIt9WQ+b6q3gzejQdM+9H525EtOVpXs1+jBymM3yAHe2Rz0GlpzzrSV4+qgU6KG+288muSobEGdh3Wmu6orVESHmAMDW9rVYgMbEEp1uIvDxeWJvXwNq8BsalhnDf+DjmfrQfl8fLv8ZomXroTsTt3ba0hWthyK0Q2gd5w3HkpauF4QZt/ra1+XYdCaP+RejaO4SIHEAXhnXM01SKY9FX7sOgTsFsc7LL3YcIY6ogRwtJhaYi3EPn4/W4kJlP+F5/KvSx/goTgC0v0j7nR/6pt6L77sYeI6uwX18gZJQTlAFIdr8NQ26GnW/5t9VbG6BkE/ExY3hhmopHfq7G5fHyUYGSwROeQ7ml2xZWF8aRQc/y3Wor6YldXDVUmNzy2Y5iluf4nsiarE7yattIDNHS1tlFgNzLFEMdUwY3C+c6soaS4Im8tb3W7y3sK7Nw/yTfPXTpwAj6hOuQSvoCfYXP4vhKyF/NrkHvcOO3BT1Zn2CtnIVXZ/DS+mL2l7dg1Mh5eW4GifGhkPANp7uNFTfaekgchDTUQ9/n0MekIibktNSLSAShfaizNPSQ+Emsz2tg2Z3ZZCca2XqiiaHxRsanhRKkVVBt6WB3SRO/lprJijMyKim4pxCYHhHQK8f8t+xYIaIVpcCMVwW7X0XA/2PvPMOjKvM2/jvTJzOTSTLpvSeQkIRA6NJ7EymCUhQUy6q76rr2soqrruuq69orKioWkCa99xZ6SCWk9zqTMn3eDyckGYJlV1f48N7XleuCmVOeOTPnPs/zL/cNiv8seeinU/HGvL7c8+UJKpvN6JQyXpjZh1h/sW789+CE3xJXc7zXDJEvX7789zcuddiJOPw0T1rauGvg9Sh6TeLdUw3Eai2EVe5DOPs11F8gBECqwDrjAxTbnoCEiSjVPhjlnphLD6GpOE7/itk4A9NwBSThnPEeJ0jg9u/LaWrrmm0cL2rgz+MT3Ih8UqI3sTqjSHpNHaYDsz4SHdf3v8JAQ1cC9hKmxchxtdTh7GCHdFVlF4lfwolPxeqT+gKwtNI+8VXUq2/pbLVv7Xc3p22heFbn4t+dhE1VaLc9RO9Zy5Ftfp7PR73Lxxd9+DzHxaRJryHP+RKh5DCumLGiQcGOv4ot/GOegfChUHKg81AumaqnQKnFyOHSVoa4TqPrbiYMSE9/CXHjoGCHGEppKr58b2guRxqUxDRJNkmLwqlol6HWeJLp7IXX9IHU1tWQ0+7F+WIt943RkObf9RPPuoL4UmFdC2Heaq6LD8OzZDusurUr1t5rGsrQFKwOMUciuRgtSgAAIABJREFUkwhMS9QS6yXgpRFr1Cf39uf266K7NE6srbDtGchaTUvCLF7dW+HWPFrXYuVCYQHvTNBS02DHW+tBcICZHnY5HagxmXsk6kwWO/XlBYR79e7pSAQE6FWdpY2X4KmW4adTcnrnTt5ZvLjrWGYbz23IYkuW+ID75ngZ43r78885aXiq5SSH6Fm+eAD/2JJLjcnMwkERzO4X1lVjL5V3GiPbHE5qjGZRIK27tVtzKdgsYlXNZeNNj/Bm7T1DqWw24+UhJ6KbacVV4YRfgas53muGyH9v2BxOGlrsaDPuQ7N2MYG1WZDzOXeN+ScKpxmh0Q713QR8HFYkR9/FPO5FVJsfgOB0fEP7I9R16ZlLqk5B1SloLsEY90SPksE+oXoaWi08MDaOplYz6QYbGfVr8dz2oajxUZsttoK3N9JeX8KF2ftot7n4doGcj47VU9nq5JZEJyPU+Xic/JZ/zvgbr+8tA0MgxcNeJrBoDcqyS12ILrGWOG8TrqA0jDIDO4Z8TbCzihaJnuUFSvpEaZA4ey6dcViROdrB1kbS9kW8FJiOc9AClJvfEP0cAaG+QPSe7H0DZK2GrU+KDyCvEFH/PWYUQlCKu04L0BY9kQ0XJaREaXqEalB7g7mZpiFP4OUZLMbYtzzmvk1oBggC8sDehHjZ+W5HHh/sE/VZPr8xnHo8MQRFUHy4hHWnK/hMp+TF612M7BXEyEgV37n3aTEqwR9vDzk+jjpkGx9wb9nPXk9g3CSGRIRS0GDlk1F24s88g7yiEHvyHCbefTtHcsrcyIeGIvF6AFalNw1tPa+vyS7FL2s5fpkdVSajn4bhf+6xHYhO9JcbPBg0Cvyr90GADAQpaPzcXOZj/LS8NLMPT3x/DqvDiYdCyms3phHm44HN5v6bLKxt7STxS9h2vobCulbSwryQSyUMj/ejb7gXZpvzR2PLJQ1tvLfnAt8eL8NPp+Sv03szIlKHInsVbHtSTFb3ngFjnhbNo7vB31OFv+d/buz8/+jCNUPkHh6/n7PHxboWPtx3kQ1nKkkM9OPhyZvot3MeNJfic/5zSJsHZhNk3C7eJC4HSBXIcjYiK9wCIf3AEIuw7zWIGioa2HZHaAYpLfuZ3/c6vjgpNl0E6VWMSvDn+R+yCfVSs6bfaXx3/a0r83/oTXEGLUhpqKvmbflSPvysCIBATxUfjFeQcP4NFIf2iQ0ek19hkvkYzhGDuHVdIY1toUxNfpa5s7ww1xSS6OUg5NhL0FKDkL+VgPLjNPf5lHv36AAn0I5d2khaYhSJl6lDulLmIZQegfRb4MSnyKtOgG1qJ4l3ovwExI4T/+20Q9VZMSSUPBsqT0PeNhyzP0Wy7xWE+jyao6exw2cu27YbmRgRR4QuGMmlDlJBgOseAp9oTteoKTlcRJR8MAOuewxF5ntiXXrfhZC/FeLGg8KD7EojaoWU+8fGUdVsxqPhPGObD7Lo/ExOV4gVMbUmC3d+eZoN9+oYZM/k3gGRHKwUcAHB3irMNgfPbSjgzbGqK7bsS5sKeT7BhlkfTe9N8ztr62XHP0BprEfZ6wn3HQRB/HO58Lmwltv6zefZ7e4x4H5hnrDu6659DrxGedhUip0G4gN0+HabzUb7anh9Vi8eXpNLq9WBt4ecN8aoCJHrYOV8cbarD4UZ74odqIjuOzf0DSEtzItak4UgLzWRBvH+uvw+s1/mfapRSBkZo0MtcS/b1Knk6H6Ea+0OJx/uK+zUlSlvaueOzzNZdWsS6evv69ow63vw8IOJL4L056nn9+SE3wJXc7zXDJEvWLDgdzlPdoWRFzZliyVUwJGLjSwsl7JuyW5i5fVgiIWSw9jjJiDb+phoAACiNsv1b2MrPIhl4EPosr8CpQZ8E0QPxVNfiLO5tAXg3wvfnB94XHqRubNvpk3uhV4Bbx2pQCoRmJuswTfn857lW8YKiBjGOVs0Hx7u0uuuMpr5+3Ep72lkKGztoA0g0xKCWeLBn7/N6dxu3dlakMjJr/HAarPxQepCoqv+Ir7Z1kAvRRXQNXuM8/cgu8VJwvVf412wBrVvOHa5B1K1NzQWQlBfMeTjEy1WnVyO7vFciRS7PgKpXxKCywm5G8EzmLdztRQqn+HmKT48t7OGsyfE0rLHdreRMOdTkiwnwViBoPWnydhEg0bHgxsudrbAB3tl8OW0UURe+EJ82Pn3AgTOVzRz71cnOuWAEwJ0TBuTRpXQzulj7nX+DqeLovpWYrx8GCyRkN8iw+aAgZEGvj4uJj699V64gvoiVJ50+0xIlUTv+hPO0U+7u9YD2vw1tAQuobQhGF+tEqPZhsErEln6rZD5CbTWMtX0DfaJS/j8aDneHnKWXhdJWPUq9+tpa+NAfhUP7yygf4QXr83t29kUI5NKmByrImlUFQ0OFQH2ckLM7bDvQ1EjHsTKoK8XwJ17OsWpZFIJcQE64gLcy+wuv8+ifLWkhnpxuqyJOb09uCfkApEFn+HaFQDD/gThg8Xr8BOobbHwXaZ7Q5PLBfm1raR3PNQ6ce5bcfWhuyxZewX8XpzwW+FqjveaIfLvvvuO2bNn/0/PkVdlYuv5qk4Sv4Q2q4MLLUpik9MBqNIlISveg+8lEgdQaDltCeJ9483kr21nTvxMpnruInj1Upj2BnhF0Bw6Cm1zHtIdz0FzKRog5dwXEJIBfvG84izn0VHj8ZJk4/RLQlLvrr2Mxg/2vEh5/28B93huZYsTyZB5ENIbp38S+fVKjK6eQv87squZmxHGxweK2G9LJFrp2anE6JR1xScNGgWTehnwdtRTb/PAFTaGkA3zkV0iq/RbQKETZ+pbHxerEiKHieWAl9D7BrGTUCKjfNgL/CMnnhv7j2LwgaUIQFtgBhuLnGRXmhBUOvy9tFApEpifTom/vA1h/ZOiVri1BS9AMuI5BKF35ykqmszct0PKl35t6MzNMPQBnDIVXx290EniALnVJg5V+nNDSG+8PZowme3IpEJnuZ+3RsFJRzoL1mZ28srO3FrevLE3gaZzKGU2MtOWkcJzKCqPgi4Q+7gXke16FgCJ6wohKLUXVa1Q1tjGU2vPcbq0iXG9A3hk6P0YwgZA3hZMSQv5al0F6eHemCw2HvjmDH8enMFdXuGdCWJjr5v4Ok8c1PHiJg4V1ncSOYDgGUREdCKRq28TSXvk410kfgnmJrHO/mdUBi+/z3w0Cl6bm8rGs5XMEXYSsEcUkhMACnfCki1i6/1PQC2XEqRX9XA00qvk7iQO4sRH4a7m+EvHeq3jao73miHyhoYrKNH9CrhcLsoa27E5nIR4qVHKpWRVNtNmdeChkNJmdV86apRSccbVXIa53Y6yxd3+Kj/jOeavb6XVKpLiCzVQ0280j/ptQXL4HXb0ewdvZzBJkhLUpgqxsqO1DtrqofwYxI5GdXEPfilzMMkicYZm4NlQCLVZkH6reAOq9NDvVsKU7jeETinjs+Em1GuWgNOOBJiTcjPnYm7v8bnDDR6dhsv5zYLYDWkxQuQwInQu3hqnRgAyvNtQF3+J9sirYuNK6s2QdAOcFsWhOPEpzjmfIWmrFVcKxnVivXj4YGguwxE5HLshgeLySvLDPXj9hIv8uno25TSyZt6LhKXlInHaGajSkl1pYvWJckbE+3H/2DiC9CqK61pxlawTq1K6VaZ4nv6I6fFv8fGJrtc0KiWO3jNpG/ZHMaTlcHKi5DIiA4rq2rAmxbBuXin20uMo7CZKVYkcbAsjMVDH3zfn9OCV709W8NTwKM62+VBjM1OQ9j6BA8ycK6kjTohmxJi/odr+OJiqeszYS/o/ya48Ca3KWnbnirXs3xwv40JNKx8vno06ejTnz5RTWNdKYTdbwX8fNTJtyrOEHH2euphZrLIPJbOsFZlEYGSCH+DC4XS5NSi5wgayYcAKZG01xHupiemuNw/irFn4+f6+K91n0X5a7hlkQPjwXfc3nHaxGepniNzhdPHQhATu+eJEZ2I2OcSTPqHekDwHsr4TCV2mgrHPiE1MvwC/NSf8r3E1x3vNEPlvCVO7jVUny/jH5lzabA6uTw3mwXHxSASBH85WsmhwBO/u6XKLHxxjIEFvh/X3w5mVRCp1WCa514PnOkMxaAUSdUrOlTdjsTv59JSRW0bOJ8ReSoiknrLSJmzhIahHPibagkWPEhN4+18FuRYG3Ily859R2kV38ZZxr+DUR6Db+RhCR+LLpQ8jdcpUbhrgycpjpbhc8KdBnoTsf8DtxpWe+ZLEiGH0Cwkis1yMwcqlAnP6hfHSJjHccl2cH+ag+Tg1gXjUnMRv891MsbaAVI59/EvI9i3r+oCZH8OwB8Uyso4ZfE1TKwalD53ac5mfQPQYCOmH1DOY0qpKJq4TkAhmlDIJEQYPyhvbKahqJEHfimTrU9w07DW2exkoa7KwJ68Wo9lGerg3Kw4XM39Uz85Fh4eB+o7caIS3kteG2klq+A6hyEGRZRJv7DEzKyOCyX2CelShDIj2oaW6kMQ9tyIYxaV+iCCQMe8bpB7JcAWTX62Hmt21Gp5df7KT5PuGeREfqCdRKkPVeypEDQJbO67q87Qa6zE11VGmiGJ1hQ+vjZDQ2lJC6jgZ3xYp2ZzfSmZJI5VN7fhXZ4L5yiV5jcEj8b1lJN+fauTFjTn465TcPTKG1SfKeXFjDjmVJm4ZEikmUi0mJA1FDAmC+7cqaciys3LkMnQ7uxLB9hGPI6srgOiRXSepK4CCbVB1DuLGiiuqH4EgkYt5iMsh/+m4b2VzO8vWZZFVZeLP4xOw2p1E+2mIMHjw+ckqGlz3Mm3m4/RXlKD29hebwv4fvzmuGSKfO3fub3asU6VN/HVdVwPMmlMVBHmpmZMegsls52RJEw9PSKCxzUasv4bhsb74Zb4Mpztqns3NKM5+SdPk9/DaK8ZGdcFxZDTUU9Fs5s4R0eRXt3CgoA7fwDCEg9/Q+8jb9BYkuPovEWuQs9fRHjWeksg5SGfvIEpWh/Sb+V0VHA4r2m0P0Tr9A4TarrEKzaVI8zcxKuFeRsT7oRAcDLYdgSOXlRci1p+/HZ5HVtIQmhSBCEpPXtlRiEQCD46MIkov8HThaJJCvLCp06HvHAbo6omU1KIq3tvzwhXtg2EPwC5RhTHb6odCFcFQjS/Y2mH0k5CzEY6+B7Xn8U9bwsMTw3E4Xfh4KDG1W8nwbiHBnovE0g4jHyYh8+982+dGClRJyH1jcGj8mf/hEQCM/v0J0fqL1wtAkFCc8gCp9hBKTBL+1q+V3ltv7tQ0jz+1nDvHfsG0T2r58vaBTEoOZNO5KqQSgXkZYZjabRjaTneSOAAuF5Kdz9Ee1J/+EaJMrMPpQiWXMKVPEHP6h/LJgSK3mfrJ0iZGJvp3JR01Ys24pOwY2h2P09LnDvRCHc8mtCDf8zZC1RligYHxM4jLWMy7J9qwOVy01ZfSGxleHuFuFUz3jYomKdQbQRAYH9POoThvUiP9eGlTTmeT08cHimhqs/HieD+UWx6GnA0YBIFPUhZwovfDPHxQwdwRX2Nw1NEo9WVHtScP9pXQ+Wg0VsLKm0StFRAd7Yc/zE1z77nyTaPSic4+K2/q9ppeXIH9CIrqWtmTV4Ovp4p5YV5syarmVGkTL96QzLz3D3eGtb7JLOeDRf0Y5//zcfHu+C054ffA1RzvryJyQRAkwFJgGTDa5XKd+2+Pde7cOYYOHfprhtOJU6VNPV5bc7IcP62Ch8bHU1TfxpnyZm7oG8ygKF/0zqaukEIHKn0HM3uHHzf2+og+USE8+G02ze3izXjoQj1Lhkbyz2nhKC681bXcdjkRjn0Io56gRN+PlyrT2bjJhExSytaZMqIv16FwWFG2ujebAHhUHOLLyimUG628O6wNmadWXN6WdWuTFwSQqQk4uYwAxNWDrf8dpI0Yhsk3hY+ON3HnmSYen9SLP6481RlKUkglfLBgPP3bqukRYfeOgNKjMOYZyu06/pEpobDRyDtjP2CwpgrVrqfEcBFA9nqUDRep9n6JT06JM+OPr/ejz777kF3y8JRIYdwygrb/lSCHFRZvolgb0elZurXak9y0D0gjFx9JO7nyXjy+X8bFhhzm9A8lsmKluzGFy0VM8dckhyzl00PFPDUlkYRAHQ6ni125NXyX2cLssQ6IHStWc9TmipeqtYZGo4mTpc08PbU3RXUt9A7W89buAlafLGd4nB9/mZDAK1u7Gne0Silxlzv2hA2CyGEEHvs7gT7RojhU1ZnOt7V5a7hz4ig0vsOQSgRqpIGkH/wDK65fy9k2HxrMTuIMCtLCDWIdtqmKiDUz+bc+jk3qxztJ/BLWnCrnj8lWInM2dH5+2enPsUX/gc15zWzKBfAGHEAj9wyMhqMfQNYa8feSPEsUtrp0DQ+8TmvgaHT6HyHnmFGw4HvI2ySuzCKHiv0MV0BJQxuLlx/jYreQ0YPj4qlrsVBQ6y5FAPDmzgKGxPiiUf5yyvktOeH3wNUc768VzUoFjgBXKGn4z5CV9ctctn8JQr17LmcjDRq2Zdfw1NosNp2tZFpKEBOSgtB7yEHhISZhLkNzu43XDxs5U2PvJPFL+OZ4GTEGNdLCHT32o6WGzc6BbMwVKzTsTheH6xQ9Y4MyJWbPyB67l4dO5kxlK/k1rayvC0K+axn2sctwBovJWNTetE19l0rDQMxJN2IPHYwzaiQNcXMo1KSyNqcVk9nG7cOiKahpccsHWB1OvjtZyWmPoTh13XxgVF4QPwVC+uIUZKhC++KQyDHbnCze1E5uq7qLxDsgrT7HMF+xQkQQINyc00XiIBLIqS8haQZM+xcEpRJh0PD5bQN5aHw8flol929vZfyucG48k876hhDy68zYnS4yixtx2dzVAAGkDisyiQS1XIpOJSfGT8v7ews5V27k/gwPfKXtYl4iIBnGPQdyD8pH/JPNRQ68PMTKkhh/HY9/f5aiujZcLtiTV8vevFqGx4mV7VKJQL9w756kow+G69+E27aJaoiXdGq6QVt+gMVJMkK9PdjeFEzRmPf419FWHlubwz+25HHnV+c4WtwREqrLh9ocNAXr8TGX9jiWXi1HWX+ZtELKXNKc59k7poRVkxyMjBR/6yOitRgyX4eND0Hxfjjwujg56b+k2/dhx9jcM7fg9n1lrxNVDDM/hs9vgD0vgrXn7Z1V3uxG4gCfHy5mfFIA8isoC1odTlyXJyh+Br8lJ/weuJrj/VUzcpfLdRJ+mTmqIAh3AHcABAYG8v777wMwYMAA/Pz8qKur4/333yc8PJzx48fz4YeivrVCoeDWW29lzZo1nS7Vs2bNoqCggNOnRdnRIUOGoNPpOtXHkgYMp0+InrPlogONRiFlXO8Alv0g3hQVzWZe2JhDoNRE5oE9CILAgpEPoi47IoYQgKCa/dw3dDQv7arAQ9Hz8wmAgAtL8CCUxtVu7zm9o1h91F0l8bXjdsZOfAv/bfeILjdyD+rHvopDHYJHv8VITnwKuGiJm8EPtv40tIo3yf7idv446gma2p18H/EygVFGaixyzl/wJDVUz6rqWwjQKZiUHEikvYXo2u0MNu2jISiVvY3pyLURqOVSNznXWpOVt8/LMA/6hJGK80haqsVk6/EPoPggEsAgU7FixkqGfCU2o6hUV4j1SqRYBCXQjlwiQW3ruRLCVEHN8Jc4V9ZM/4snkGPH3ODgRh8HXpYyxkxXsdMYwtP7Wojx07JseiLbs2uJ8tVijp6LJvsbt8MVRs0jZ7OROf1D2btnNwYvHWv+MIiKikquK3sH6Y7PxA0rTkLhbipnreH5YxI2ZYkVSBIBnp2e5NZgA3DkYgP3jY7lYl0rD4+NIjlEz4oVK2hrE0ns1ltvZe/evRQWirmVKZMnExQ1EknHrL8T3uG05O5jX62OGwaP4dDFALZd6FqoJgfruVjfyp6zF0nRarlkyNa75QipQcM4XdlVc/7n0VEEOboazuh9PbTUoFl7GxogDHhhyHPcZEzmoXQB6Q/datMBGi+CR5f1mCVpHkdyK4kfzBXvp0BXFRGXmpQu4diH1IZNQp84nOXLlwMQHBZOq2/PWLex3UaYl5oILwUfXtaNfMuAYASHlfffF4/v6enJvHnzWLlyJUaj+GCbP38+x44dIy9PDAdZLBZyc3PZs0fUWE9KSiI5OZmvvxY/p4+PD7Nnz/7J72nChAmYTCYOHhQb5VJTU4mNjWXVqlUA+Pv7M2PGDJYvX47VKt6zt99+O1u3bqWkRKwqmjJlCrW1tRw9ehSAfv36ERoaytq1a8XrERzM1KlTOzlMIpFw++23s2HDBioqOoxqrr+esrIyMjNFraVLvPfDDz8A/GLe+zEIP/eUFARhCxBwhbeedrlc6zq2KQKm/tLQSv/+/V3Hj7ur6eXm5pKQ0HNW/N+iqtlMTqURs82BQi5aWHWPU8b4a1h91xD0Ht0CDDXZNJecRYYDTdUx6tvsOJPn0GgyMXeLzE2p7Ynr9Cwte5zKYX8jaPNSMIkhEkvsJMx9l/DiGS0rT7vPfl6ekcBs5VEkTUVibXZdAZz8DMecT6G9mTKbjqcP2dlT2FUH/dS4MG4zvQcF22gOHcHRwAU8fsDKncPCeX5TQed2w6N1fOj3DYrTn3e+ZvVP5avYV9hXKbhJmD4+OZFXtuTx2igF47wrUay/B0Y94W4SADiD+pI79J+0NjfQuz0TdUM2wvk1Xe8PvJsy7wy+KTPw5vFWVk2V0m/7ZXHCjKXirPDYR3D8Q7Eef8JLCJsf7lzyu/QRFE9ZQYMsgN6SEmT2dmTmBiwX9qIITMRVsAOXy0lt71tZ3xSFXaKgT4ieYXHdekPr8uHtgT08QrdNO8LSb93LPB+ekMDLW9wJ2NtDzmdLBhCoV+H3Y50vl6P8JKy5szOEQ8xose4+dgz0miaaGmRV8fwP4kPk+rRgpBKBtacqcLpcTEkO5KHQbCJ3iXHr8ozHOakfSx3eJAbrSQ3Vo24ugC9vEuv6Rz0Ou9w9S5GrKZn+Hf7WUlSb7hdb5rvb4U19DTI/hZQbodd0cqvbetxnLpcLs92JuvoEfDi25+dcsrkzVl7Z1M6/duQT4q3m3zsK3CRobxsWxf1jY9Eo5GSWNPLZwSLqW60sGhzBkBhfPP9Dy7bfmhP+1/g9xisIQqbL5epRRvSzM3KXyzXhfzOkywYi+5mh2NrF7HtjkaiCF5jyowazAIF6FYF68YbMqzYhl4jdbiHeapparVwX5+dO4gD+vTjb7ItQfpyhxz/CMGApbLkLv/ZGvpy5gfW5reQ1S5gRYWNIzedQeZqGhnqqJnxHtFBJbr2N1SUaJjZKWZCiY8eFNmpbxFrnlGAtA5s3IrF36Iec+RqHdwy1E94j8PDbEDEUuTYJH1VXtGtCgjfjOSwmqgB9zjeMqjvLq9M/57GN7kvxudE2FAdWuL2mqDlN/9RqPPv0paShDYfTxT1pcrQ6canr6TLRjhJn4ixUlzW7AEjqcok1ZyGv3g95m6mY+T2yyOn4mrKRKNRISg4TfuQd7oyfBUPvoUZroGnqh3jtfw5aa3ElzUSQSKF4P5WCL5ZhLxPkrEJ5aoUb4QrNxYSZThGZuwnyNkL6IijYjtJYAYIEIfI6ivvcy/7WMGyCnH7h3j39LCVSkMh6EHmTuedEJafKxMh4P3bndcnf/nV6kmjr1g21JgvlTW1olWKoRCV3b4yp1sSjH/JnVE0XxNhSxSlRCXDoHwE4UdxIm9WBTCIglQhEGDx4Y0fXw3fD2Sri/Psxcs4BGpqa8dZpCBAUTO3TrYXdLxEWbxQfFq3ueuIA2NoJbzhEgSKB40M3YmxtI93TSMrpZSgkQNxEsbRULt4Lsst6F/KqTXx9rJTDhfVM7O3H9JFvELH7j10b+MS4tdTvyq1h5bFSAjyVPDY5kR/OVFLZbOamAWHMTA9FpxLvqYxIH/qFe+NyuZBK/7sI7s9ywjWGqznea+ZK7dixg5iYmCu/6XLBmW+he7tvxu2iUJPq5z39Qr3UvD4vjWfWZvH9yXICPJWM7uWP0+kUwwp1edBYDH4JxBsSyWqLp7n3fPRqb7EBA+hVs5FeZd+KmssHz3XWPiuVSr7Md5EQEIuHh4wb+hjJ2L0AiamC7zMeI18Wj8wzgISm/fgffEb8LJ4hEDGEuoipNFjAmPQndpY40Jpc/DW5kjvCLTidTiL9pWi++4vbZ5HVZROlbEEtdw/3SHD1bL4AgnQKtta0sXhIBBnKUmK3L6Zh9CssmxxLja0J7Yln2NL774zxruZyFY32uKnUqOIJVZ9HOuJhso1KZCYVIw694dblqMtbxd03LYLMN9HUn4ao4TgSpiE99gHmhnI2ec5n2YlqmtqsTEtK4cGEaCIq7nU7l721HmnZkZ5fnssJF/cQcXEPETM/EGeWl8HhdCH1isA19AGEPS91ve4TQ7BB10O9r9Zk5vnrk7llaCQNLVaifDUkBbv/jnKrjRwsqKekoQ1ju40hMQaGxfkS4NkVYtqeU0OSPIx4lQlV/kaavHphm/EwAR3EV2U0883xUh6dlMievFpyq3oaR2/PqedkmYLduU0oZUb+dkMyGZdv5Bkk/tXliyWC1m6dqyH9uKBJ5abNEmpbxG5gQYDlc5czIlQmxvW7oft9VtXcztJPj1HcIIZzsiqMnIhP4N+jnkd79HWIGiHKJmi7FuQbz4rnqDZaWLbhPMNifbl5QBj3jO6pQy5qpv/3Tjw/yQnXIK7meH9VslMQBG9BEJ4E9MAdgiAM+m2GdRkaLsLmR9xfO/ahaCv1M2hut7LpXCWPrDpDQYdFW7XRwutbsjBl78C5/n7Y/5rYGbfpYRw5m2lBRXX/v2AP7nZL5W2BqOGivkgHibeHDOGrIg29AnUcyKtEp5Ihb76ApLEQ7GZCDz3DqH03cV3lcvzPvNNFtMZyOPst3tYqAgMCeXBnKy/trePJHfXM2u2NJWwY5Z5pWFxX+HoEAaXdxINPkWstAAAgAElEQVT93JepWyvVOBKnub3m9I7m62IP/rU9n8e+z+LeHTZKJ3+G15FXiDaoyWzWUd7/EVLbDnFK6IVx3KtitQJgjRrDDr+FtKiDqEy7H+e5NRj0nkR4KXq0qgNoKg+jyVstCo2dXIG0NgtC0jib9DAPrC+lodWK0wVrzzXwQU0Cdr9kt/3z5b0oH/4y5UOew9pUKc4iu0OhFROY3VDZ3M4Xh4uZ/8Fh3t9fxD6fGeSNeAtT4lwqBv+VwrEf87cthTw/o0+n2NPQGAPPXZ9MTICOUQn+zOoXSnqEN8pus22L3cHJkib+tSOfTw4UsepEOX/+9gz7C7qaxMxWO2PkWaRtnYvHzieROC20+aezPK/rcZgSqqessZ3PDhYzKz2Evlfwv0wK9qS4vq3jvE6eW3+ekvofcdnxjYMFq0XhMLlabOAa9STHreGdKz8Qf2av7q+jRXtl87xLKKhp6STxS9iVV09xzAK46yDMeAcC3GPhGZFdq2CnC/bm1/GfpTD/H/8L/NpkZyPwfMffr0J8fPyPv2kxXVnr4woiR5cju8LExfo2N/F6gOf7taP/bn4XuRbuEkvljjxHbsS7PHKqlW339COkQziKqjPgHYFr8j+x1RdTo47mOL0Y6OXPoOJ3udl2nKK6hWj9InsOIn8r1vSlKPZ0a8CRSJHJpPh8O5OPEhaxP2kEYdJ6EpTFlJQ3sqPYizU2BS8nzEab+13nbs3Jt+B0CYyUnuGLyfFsrVDi7wGjgizYfO7G5dcbWeF2nMHpHPC+nn9u6JoF5tS2c6Qlkurg59E2C3x1rITvT0q5NS2F4XYL9+f1YWTaF+hldjaXyght88W73U4I1bgSJqNpymF3gycLfRORdqtOcQWnI3hHia32DiukzgO7jZqQCVS3+fOHkQbWnqqgvKmdWH8tft5eHElcTmz5GvwK19A26AHCZQ6aGls5L0vhgDKDJZ5NRE16GQp2gk+UeMxupGK1O6gszGJw21lGRLfh9O7Dok02ypt9CPeZT12uBd2pBsb1CqBvuJ4N9w2j1WInwFOFRinDbLOTV91CnclCiLcHBq2CvGoTpnYbwV5qGlssnTkVQQBPlZx3d19gbKI/Joudg3lVnL3oRf+MzxjUvIWAs+8QWnaMObM2d44xNdSLV+ak8sOZCsoa23G5IM5fS36NOKEI0qsYkeDHymNdYTKj2U59q5Vww4+0sYcPFMncYhQlHWRKjBUXemxW32p1i19bbA5kUonbfSa7QshDEEAqk7qpKXbHlJRg1pyq6KxYifLVMPFyk4srwWISS1uL9ol5hMjrwDf2J3f5SU64BnE1x3vNhFYyMnosKLugDxVjhd1n4HK1m0fhj6HSaMbhcLlVbsT6a8Ua5ctDEYW7wTOMQA9Rf+VwmYVZo5+CxCmiM7xPDIIhBrtEi6XFSa/6iyQc+gOUHweHjejyY9jmrMAVPwkhb1PnYZvDRnNYOZL+w5UYsj7BqQ1EkjgFyaE3wVhBwMVVzEgC6W4xLNAHeLT3Qh6zzOIN6UKWTpmEsyaHeo9Y1tSFskSmQK1QMFRWydAkNXani4seKbRKrehNVaA2UO87gHs2tfUwGCg32bnYrmO8xMjSgf68f7iGd44ZeeeYkX/MTqbN4iCzxsSUdAPNFgdlLXaia06iP/cFa8PeYJh/O66Bd0HZEag4iSN6DFKND2x/FuauoN0lJ1cSTaFRSl21lc3nysmuNHHXCLEUMkCv4s1dBdgcLoL16bw2cx4Dts5EqM/DEwjXh+E7fjV7GgOwhPcjtv/SKxKOpSqPtJ2LunTYJVLeGfMp036QdppjN7fbmJoSRK8g9w7Sdqud5QeL+PtmMUk5Kz2EpjYbOzqMh5UyCf+YnYJSJmFKnIolEXX4tuTSposEUySPrq/snJ1/DszrM5lnws+jLtlDkKSrckejlDG7XyjXxflyprSJO7/IZF7/cKakBOFyie47eZXu4RYfjQL/jmSr0+mitFEskQzxViO/dB1Unm4hxfQIby7Xplo8NBIfjZI6k4Xt2dV8dbSEaD8NCwf26dwmzk9L/wgvjhd3jfmmjPBOpcQrIdZfy5dLB5JfLV7jOH8tQb/Ejf7c9+6hUa9IuGWd2LvwI/hJTrgGcTXHe80Q+RdffPHjouwaA8z6EH54SLQN846C6W+IS82fQZi3mle25HLniGhe354PiDFVmfwKusoSKS3ho/g2u6N5RibQghqzMgzdhb0o9/4DV3A68rRFNLji2Zxro5inmD9aQpS0GmVjHkGmclwByRCUIjal+MRQqx/M+iwJbzf0JdU/nT9Fl2PYdFdXYq7X9Uj3v+o2FO/zn7No5BQWb4c2eQL7CgyUNrbxzRQzgavndVYmuIJSyRnwd5afNBNhUHOgYhYTIiV4WLTM6Svho0NdXY6CADKJhAkJXvS7+B6Dw/szfl4M9S49njI7cdaTaB0lnI4dz/snKtnRUekS4RPLu2NeI9okY9DhJWCqElUi/RIRqs+KS31TOa6yTDYLo3hwa1YnqdwyJBK708XrO/J5fW4af1rZVU5X0Wzmha0X+cw/HX19Hqj0HBjwFnd/dxFjux2FVMKz1ycxs2+IW+gDQFV51N1Mw+kgNutfTIx/hg05YjlbqLea8CuQUn5NSyeJA4QbNKw6kdf5f4vdyVu7ClgyJJwl9m/w290l1+AoGU+E+l66SYfx9TkTt4ybS6+KI6h9QnucL0CroE+IJ6Pi/fnyaJeJhyDAB4v6d2r/eHnIeeOmvoR4q6lvtbDicDFv77qAw+li0eAI7hge05nA744+IV4svzWDV7flUd9qZfHQSKalirHxb46XdlbonC5rZvO5ar6/ZwiJgZ4YdEpem9uXffl1nC5tZHCML4NjDKjkP00LQXo1Qfr/wA3IWAHbn3Z/ralIlDr+CSL/SU64BnE1x3vNEPnPIrAPzP8WWmpFf0KN78/vA/QK8uTWIZF8f7KMV+akUNVsxgWUeU4l/uxnXWQqCDiTZrK+KoTTFSYWpeq4TpbNjpM+jC9YhrJ4p7hZYxHyor1cSP2Ujw43Mz01mF11cu7JbEYmSeXB4YHMqHwNr4sbwMMXznxDcPICDB5LCPT04ZODRSz2NWHoXl0hlblbmnWgn0cNW0a0YwkJIk0rob/aRLhQJIYZjn0kdpJWnibEUsjx4mBCvEM4XNLC4RIAI9/elk6jxcXGs5X46ZTMHxjB6hNlPDoiAP/C1eTH3cDzu5s5VV6BQirhsbGx3KyqI79FyY7sruV6cYOZj7KV3J9qFkkcROeh+gIxyRIxRNxOm8ITG5rdZoafHSrigbHxnClrpvkKRruny0009uqPnpVUJt/F/bttGNtFTRmrw8nj358lJVRP0mVekjJzz7CavK2aYG8xuaZXy/nnnFSxlNBYgaXsFNZWI0bPGErM7glAm8PZ41gFta38+4Zo/D570+116YWtTB2+kC/Odj1YXC6wCQqcN7yHxNAtXGC3QslhOPIugXYzjw9/GZlEYGt2Nf46Jc9NT2Z4rC8b/3gd9a1WAjyVnX6VxwobKK2s5ra+Wr7Ps/LxgSKifDUsHBzZY6wyiUBsgI63F6SL7jwacZJS1Wzm3T3uYZd2m4NTpU0kdhg7h/l4cPPAcG4eGN7juL8ZHPYrh0avkGv5f/x3uGaI3NPz56tPLl9S/hJolDJuHRrJyAQ/TGY7KR2NQrtM4QTMXYcufw0SWwvEjUdw2LghsIaMhVGEW/I5V6NE7nCh7iDxTrTWES+tQKvUE+bjwVu7ukrK/rq1jKDxi5hQuE6sLZfKqfPNYKBaQaNDzbLrkzjbpidCF4qktRLkHhh1MWh8E5DWdattVnqiNF4k8sBruCa9TMKJF7ukSw2xMPRPYpIWsNUVcmPGACw29zBKsDkPiVPNwkGRNLVZ+feOfKRSgRitlbYbPuW1gxJOlYuxTl+tjCHyfJTlR6jw6JmzPlhq4aEBPUWuRMU98bxN8kDarO5uMy4XnbFaH03POuJ4fw36JrEjrk4V3sNB3eWCyiZzDyIXwgf2OJY9/TZuTExjdD8LIV5qUQq2qRTnN7egrMhECeikCipmHXSzz1PJeuptT0jwIsx40l1hsANaqQ3o2mdgpBd1ugg+qVdzk92Jh6IjBFJ2DD6f3hnziLnQj3/dso3qySNRK6SdrjiRvhoifbvFxB02Uq2ZjDW+hKy6lltSbmGFeQirTpQzNyNM9E0FsFuxVGbRWHGBYrOGFRc0DEuKYFpKMDq1HIkEFLKeYalak4XKpvZfFhL5LaAPgYw74dAbXa/J1RCQ9JO7/SJOuIZwNcf7a1v0fzPMmzfvf3ZseYfIfnqEN/GBnszqF8adI+PR+wYiKT0IDguUHET4/g5Ua24j9vvJKGzNNHmEY3HJxPrky2ATFPQL9+LQhboe720ul3e2/BcOf52bDwbxh1WFPLEmi+c2nEfmG03D9E+wT3gZ58C7yVWlcCT9H9ijRoMgYAtMo3bqcipsGghKRSja564/XV8gkkOHWl2VtjcSQcKFGvd4a5vZzN0B2bQ217Ejp4b+oWo+nygnTFrPF2W+7M7vqsJ4ZoiSBFkV2Iykanp2aA6P1uFTuQ9H2kL3N9Lmw8V9MPgeAs35BF5m2aWUSZAIsGBgOMeLG7ltaGTne55qGX+dmoh3ZF9Qe+Nrq8JP6x7yEgSuGE4gOB3mfQm+8eLqbNSTyNLmEuuvZVC0oUvPu+IEkorMrv0cVvqceIbnpvfCUyV+r8eLG1h2fRK6jpb8wRFaHowqQZ39rRg26g6NL6GxqSwaHEF8gJZ5GWH0izSw5Ltilm3MoaCmW2ngudU98jCq/S8S4a38aWuzipMEbViArOYsmKoIOPZ3FqoOMCzWtytO7nLB+e9RfjyKwE23MWTXPP7qt5v1xy9wvFhcrSikEu4a4Z5H8tUqsNidbrK6/3NIpDDoLlEywRAL8ZNg0boeFTGX43/JCf8LXM3xXjMz8pUrV/7+F6KpVPSXTJ5NTWMz5cPew5NWIs+/jXTvP0iZ8RVvV/kwYcAf8TjcFcN2hg2mShnFmF6eZBY3caLEnfiiDR5Q30r+6HfZak+nrKlrxm62Ofn2eAkjPT9DdlZs4NEND2eDLYPVsr+QPuhBztTBqpUm5iUP5unwZuQXr6DnYiwHfSiVMTfy3gUfpg/0QC4VOFXWTFljO2lhXhj1ofSt3cGzTR9yf+podE25qCXjONSUxqGLTSQE6DjZITA20GCGdc+DtYV0WTDzU27gizPigyE+QEtciC8LsgaxOGko4+eMx9FSi8Uzgtw2PQ0GG4OkOQTuf5K3Rr7H/ftUlDaa8dUqeHRiIk3tNvbl1XGytImnpvTiwXHx2BxOrHYn285XkzbpZjzixxPkcvJarIK7VpygxWJHJhF4dnoScR2u6tTmijog1VmQOAVL2BDki7cgcdpAF4DL5aLOZEYlk6K71EXYUisKSEWPFMkvdxOqsgNIQ+r47u4hmG0OAjxVBHiqGBbri7kql7Azb6DdJbZfM+JRUb6gaJ9oEDL8L3gHxzJPaKaq2cz+grrOiqgYPw3tNgctZhtalVys4lFoRa0ZjT9c3I1LqqSkoZ0Ko4VgL7W73+clVJzs8QDwz/6URTcv7JLDaLgIG9w9Rv0z/8kdwweyK6eWUYkBFNa24qGQ8vCEBLIqjPjqlBg0Ct7eXcCYxCuLYf3PoA8RV5H9bhV1yWVX9v7sjqvCCb8CV3O81wyRX9Jb+F3hYQCpnLM+47j7YC1lje0opHqeGPYGc8pfxMdSQax/Mn8vHsVTUyKQ1ZwTu9zU3uRUyHhn73memNKLnTk1mCziEtxXq6B3uD8lkR/y5x0WEgJ7Cj+VNJhxeHQt2aNzPyQodQhv7mrku27bfXnWyO0TE4mIskKNu3hSS+R41moW8v4ZB38ZG015VQ2fnWhgaKwvfjoluVUmthS00qf/HSiCUvG3tUF0GtTl8u88C0eKmnhySi8u1LZgNNuRt1Z1Npr4ZX3ME+GF3DzhJhr8BrI228gLG3O4sbcHoytWIN3+LVJAIZEROf1T/nJQw10jxiD0iUZvdrB0SCh+XjqxaeSHbDfBsXabk1e3dSUWtUoZUf56BscYiAvQMUwPr4z3RR8YgY9GQbSfVpyFNhbDitnQ3JEsPPsNrQMf5SNmMCU1GC9HO98cL+WLwyXE+mu4f1w8FpuDQX5JKPThsO9VcWaYejNthiQs+BDnr+1UIST/LFGWFmipgPwuGQJ2vwgJU2DJdtD6dXZI+nbo0lc0i9/v0uuiaWyzcveKE0QYPHh0YiIDUuYg6INFRUJjGSRMxtL/Lsa+vg+bw4WnSsY7C/oxNPayfI+qZwjLpfHH37vb6+ZG6K6mKUhoTpyH3CeUof4+2B1OpBKBr4+VkhCgI6fKSGOhjYZWK9enBRPrfwXt8d8DV/hsP4arwgm/AldzvNcMkV8V+MbTPONzntzR0DmrsjqcPLOniZCb/s3pcjN78ys5XWZiXkI8vdr3wfZnKBn0HB8dEAnl3zvyuW9MLHKpBI1CgkYhpc0Oh80+nCnPZWJyUI/Tzk7xQXeqa5atqD5FgLNn+7VMIuCUKKgPn4RHqwl11kqQKnAM/iPZ6nR8QjxYF1uK8thTqOqzmZQ0i+/tg3l5ZwtyqcDioQNQSJth/R+7Yr1KHdP6DuZgoYvXtuWxYFAEMqmAoHSPTXuU7Cap+gTVM1dR3WAm3MeD+5PaUa79tmsjpx3fXY/w1g1ruP27gk5SA3EF8uXSgfjpFDS325AIsGhwJCdL3fVnIg0eHLhQx6qTZXy2eAB6DwUlWZncMbSf+8WoPtdF4h3wOfEvAtOH8tjqeobEGHhnTyFKmYTxSYEs/fQ4UonA1sFZ+HW42uNwwolPkUx7h7ExwSKJGytg9Z1Q1KHPPuZp0YqsO0mmzAHvMCx2B3WNYsu+v07FW/PTuf/rk3ipFVQ0tfFDR9djQ6uVhR8d5eTtBjTbnha7UwGy1+NATWrwAo6Xig/QB74+xbr7hhLYrWOU0P7gGSqSP4AgYB35FHappstx1TNE/DOKlTsFI9/kkbMhZJ4sQi4t5k9j4pjdLxR/TyVN7TZmpofS0GolWK9iXFIAOtV/pnvy/7i2cc0Q+fz583//k8pVNPgP5nTFgR5vna6yEiBt5eH+EjYbPGkwtcI5cb7sEqQ4O5a0cQE6GlqtLD9YhNnmpHeQJy/OTOZiR+h8b34t942OZcXhYtqsDhYNjmBsjAqOuMeze8sriTJ4cbG+q3Fp0aBwVtS2sfNwCw+NfpjU1LuwuiSYVf488k0296e50B9b2FmKGFT3IgtTl9A4bDFjeocwINIHnHqc6bcgOf6ReFCLiVGOAywaNIUvjpbz/t5CFg2OgOA0sVGjqbhrUCMfI0Dl4IMbQrB4BKG/+EPPa9hchq2t+YrdfRazha+XpFNsdGBst3KmzMjpblrxSpmEO0fE8MDXp7A7RWs+vYeChVcysb1MQ+XSa3KJi77h3nzRUdY3KTmIFYdLMJrtjInzwu/iOvd9JDKUJfvw6jNH/H/FqS4SBzj4bxj7rFg62lINfeZA2CAu1Lbwxo58tmZVc+/oWAI9lbiAV2an4KGQM/0t99+Q1eHEVZvXReId0OSuYubARXR4PlNjslBnsroTuSEWblmLvfgoLc115ErjeGajjOiA0zw2uZcY/9cFwo2fweql2JTevFsaQWaZqPZpc7h4ZWseqWFe/HVaEvvy6zhUWM+QGF8GRHgS7vPLPDOvNq4KJ/wKXM3xXjNEfuzYMUaOHPm7n1evVRPjp+lhHJseIGHErtuQNBcx0Lc3ranLRNsrWxuhBV+yIGUZy08ZGZXgzytbu6pNzlca+ffOAv48LhaDRsHhwgYKalqYnhZCnxBPBFy8c6yeJ2d9g9fBl5A0XoC4cQQbT/HRqKHsbgrmXJ2LCF8t2VUmDhc28vTU3hQ0tPHeQRNapYyZ6VbSwvQkSM66K90BurOfcc8ti/Hyl4qarRI5ZZFzCAwbiqK1ErsgA30ffMpVrL93GAa5FV+5BanWDxasEpui6gtEk4GQDNAYUAEqwOUTI85iu8VlLWHXcbRWzvTUYN7bK0qHquQSPhivYnDeC8j2ZGJInk1Z2FRu3XaB2f1CGZngh8PlQiGVkFdtwu50oZZLCHeVw+53URfuwtl7BsbwcbR7BIvVFQFJom1et6RvQ/JiPj3vJDHYhsFDgbHdToi3mjWnxFlqUZOVpsg0vCpOgMaPrIwXOWsNQFD7YD5ehtnmZHJgIGHdL2B7I2x5DP54plOnpNVqZ9mGE+zOrWXpddF8e7yUoo62eplE4MNb+mPQKKi5rOLGqbxCFYMumNKWLv0Rf50SX21PE20MsWwoVvHozrMdJg1mcmqrSAv34o7hHQnM0P6wZCsNxna2fdBTriK/uoXr4vyYNyCceQPE8sLdu3cT5T+y5/n+Q9S3WKhvtWLQKDBofz7e/d/ganHCf4urOd5rpmrlkgbx7w0fjZIXZ/ZB281EYPHgMPqdfxFJcxEAsrrz6Lf8CcewBwGQV53kTvV2lo32uaJY/s6cGuxWM+/Oiub2ft5EGjT0UjeRLKtAJZNQ12JlXUM4p1KfhviJUHwQDvyL6A03siT3Lp4bqsBLbmd/fh1z+4dR2tDGa9vzOVPWzMEL9Tyy6iyjEgOw9fT4AbkafVspfDIZji+H1josTgmOEytg6xPItjyC76HnGR3iJMhWTMCGW5C+MxDW3StWEWbcDinz4NRX8NU8cYaavx3aGhACkmmZ+p5IqIA1eAAH4/5CsNDA3N5qJiX5IQjwwkhPrjt8O7LTK6AmG3YuI/jUG7w+J5lgLzHGXFLfjs3h4uuO9vQv50WiXbsYdr8AJYeQbH4E+c6n+fv6TNacLMfiGYl1/hps6bdBaH9KhjzPx47JZFe3caHWxMMTExEEKKxtITlEJNALtW2cDbgetP6cHPIWM7dreXSniUd+KOaf2/KwO13847gTc8Qo92uYPMetT6G4rpXdubVIBPBQSDtJHETTkFe35vHyrBT3Q4R4QlCqu1WaIKF51AuszBL391TJeH1uGoE/0lyzO6+2h9POD2cqsdq7rU60fuh8g0kN7Rl7vpLBym9xnx272MDMdw4y/rW9zHn3IMeLflwq42JdC99llvLO7gKOXKzHbLvCyupHcLU44b/F1RzvNTMjv5oYEGVgw33DKGlow8tDjn/LeTxXrnLfqLkMqT5EJLryTIIkTSxUH2S3rKehbWKgjqr6ZsbVfEJGSBgOgwkpLqg0YQ24nsK6NubpVBxr8EUZMoeEwn2dX4QjIAXt8beYmLyU+mGR/F975x0eVZn2/8+TmUwyk957JwmEQCghQECKIIiIDRQ7eZUiuq7r79VVV3fXd3fVd93GusprYRGVqiCiwAoCCihKiXQQCDWFhJA26cnMPL8/TkgymYQEApmZ3fO5rlwwZ87M+c4p93nO/dwlPcaf36+1nug0WyRHzxtJjEvD7J+AprRV0kf6I4gfF4FvFKx9Clw0hAqDEgsvBCdH/Z1FhfFs/CyfgaE6fpY8l9S6VyHhRvKN9ew6dQxLZRX9A8eRmPMC5O2CUc9A6SkYOpv6uPGcuOljzpdVs6NYR3wV3GnYj9/37zGvoZLCcZMJ9yy06Sb0U9BEXvnXiebiThmx/tybEYm/h44IPz0p2iOINkXQPHLWcvPILB5bsY9Iv+F8vk/LjlO3cXv/WWhddeSZjLw8JYbRSUFE+On5ZM5wDhVUMD0gihdXH6LQWMdpTSzVI1aw7pSZelPLPISx1sT58lp+OFtO7rTXSLw4BWouQsodyu/VKjfJ3NIadpwswVuvpb7RYtWg4xL55bUkhniyfPYwjp43EuTpRlq0L95+Bpi2EM7vg9oKCErGEJTKx2HKaDbcV0+0f8fp8AOifPlsX4HVsqHxAS1x5E0YdFqemZjMgfyK5vowE/uGkBbV9YnFrpJbWsOsj/Y0b+fUxRpmf5TNFz8bSUSbG8e5kmpmLNzFuVaFueY/MIhb+tnOG6l0D4cx5OPGjbPr9lsnZZjy2onx1bopmXr974WKAsjZDBW5pKbM4M4+U1l9VIn4MOg03J0eRbAlD8rPwc63W1JHooYSHHcXs26I5+mP91FvsqDTuPCHycsYJXdx0mso209X4edSTa9qb97YfIxfTky2qYMN4KHT8sDHp3ku468ME4fwrDqLa9xwvA4vVepWX2rj9sN8DCOeBqCi93SeOxLDnlxldv1LYz17CtxYNeVV6iousqe4imoXb7TuETy+y8IbI+bR5+uZcHAVxN0AFbkEeAbgf2YhIYZEBvqYCTu6CAbPgGPrcANicrcpTXxb0RiWznun/CmuapnV33WmlN2ny/jHFqU5wb9uk/Rp57hYmsqgHi+q5KOdZ5ES/rLpFFoXQZS/gawRscQFKREY6bH+pMf6A7D68UwuVNbxmzWHMei01DTYJvaU1jQQYNCh8YuCOduVmvfe4c2RKQA7Tl7k3W2neGREHPM2ncC3neYI92VEEe6rJ8LPwLD4AC5W1tPcNMs7XPlrwhVIDNHReXEJGJsczMof8ziUr+y3KD890wbblgAA6B/py+dPjFBCDt00JAZ74edh+8TW3essr6zGqkELKJO7uWU1Nob8YH6FlREHeGXdUYbG+XfJHWNvm3Cl2FOvwxhyk8n2QrMX2uBkpQ7z9j+3LLzx10rvwvJcwKJMhgGBRz7gt4lV3HLPcxy4CEkhXvxmzSE+v00Lx9Y3f7w6YiR7Ev+bz/e6oNUaefqmJN7eepLymkZeWnea9x66iRmLfmxe/9kJisvmra9z+P0dqfy/j/c3v+fppiXKX09JdQO//LoBD10vfA0pxBbreN/NhC7ldtj6urKyIUBxE2TMItd9MHv2Ws8FXKxq4GhDJL/cUo+xrgQowVUjeGFSH74sdqGPV6jSdcbUoLhedAZExsT8g3gAAB87SURBVEzCP7y9pS62po3BqCyE4D6KWwWo8U8h+7RyfF0E3D80hkBPHT4GVx4dGct720+z9rw3SSH90BQdbP4aY/I0VpxyBerx0btahVabLJLTF6upazDTaFbMfeviWsHe7pgsksMFRlyEYPaoeJt4/9RwH+IDPZRRsab9cLz9uRVcqKxny08XeGZCMu6uLvzv1H68uSWHkqoG7s+I4L6MaIQQlNc0sPZAAW9szsFFCH4xPpFb+oXirW/HBdYFYgI9WJg1hBNFVVgskoRgT8Ivk40ZHeDRcdXEJrp7nfkZdDb13TUuAj+D7Q2uttG29EF5jXVVxsvhSDahK9hTr8MY8q1btzpOWyedh5K8kDheGVXXVcLej5RHZFDcK0HJzS2+fE+vozLoIf6xpYZfT+5NWU0jZ8ohyjNEiZs15rMt8XnmrjcCygjFoNPw+Jhe/HnjMRrMFo4UWhvY7TkXuWdQOB//WMD7357m9an9OZRfQYCnjvEpIWw8XNi8bnWDmeqGWqL9dJA+E3a+AaY6JX0+eRKapdPAJwq3cdPRuhTbVESskTqMdS0nYaNZ8s2xCwwK81L84ZciPLwjlH8j02H2N8rvdzUo61wqYQvw4yK46Q9KCF/JCbwi+zFJ68E7O2t4Ymwv1h0435xZGOmn58lxiczbdJwx0/7GkPofMJ/+lrzgMXxe1Yet+6u4KSWYG4Jq2Ti5htpGye6aUP53RxU39g6ioraRrIW7cNW6MPOGeDJi/ZpdDwEeOsb1Dmb9oUJ+KjTy2Oh4Vu/NJznEi/uHRuMiIDHYq93qipe4OTWELT9d4EBeBQeaokKenZDE50+MoN5kYeuGL4jwUyoKfptzkZc+a2nA+/ynB/Hz0DGxbxfKvAImswUJLdmbQLCXe3M1xGtBd6+zuEAPnru5N6/9q8UN9sKk3sQH2t4Ik0O9cNUIq/6oMzJjCeni73Eom9AF7KnXYQy5w+HurUxUmRrg0zYVzX78EDJ/rhgyjSum8b8nwmRi8USIjJSUjoom0KsUS8qduFQVUjnqZf6+2XoUUtNgprKuEYNOg9kicWnTwPqHU6VMvyOBgdEpnCur50JlPXUmM8cLjTycGcuopGDebKqMd4kpA6L47U+e3D/qbfoVfAIaN6UBh8UMZWeILdnGY5k38ua3LX7XsQmeVNXaJi2V1zaSGeuPDHyURn0Q5sgM9C6tDF5gYkv1SSmVgmbr/luJeEmcqOy7ZdNB54HL4U+5L/M18lP6U1lnskoPzyurpaiijjBvd3JFGEMyn2RDeS969RtB/IUqFqe6MsSjCLelt+JdqehO9Y0hfdp7FLhGMmdxy1PM1uPFLJs1jGHxSsNhN1cNT9+URH5FLRsOFxHu487rd/XnQEEFTyzdi9kiCfV2592HB9Pfps1bHdtPXGTtgfPc0i8UX4OON7fkkBHnz+T+4fg3uQaMFS2j/BWt6opfYvXefBtDfq6khoP55dQ2mkkO8SIx2JPdZ8v457enaTRZeGRkHMPiA/Bwc7zL081Vw4PDYhgS58/58lrCffWKwW6npkvfMG8+emQof954jPzyWu7PiGba4MimzkEq1xKHOVP69r18AZ1rSk2pUtu8oUaJ2fWP7XjdRlsjh7kBglNg9HMQOxJRdo7e5lz2iFjuXnqG5bd5kbDuYaXrEGARnjRa7rb5Gn9XE2PjPRiTGsvFKuvQNY2LIEgvyFpxlJtTQ4kN9MAiYfOxYrIuVDEo2o8Vs4exYncuxrpGJqSEUt1gwowLDy3LYemD95GyfLhVWzDXY19wT+ZQBobFcehCIy5uBsYHlFJSUdBWGtMGRVCn8eC/DqSy+1wFyaGneX2ad/sZgULQGD2K89O/wtXSQJi/t5KGPfmvsHoOWMzE7vodr8/YyM832VbByymu4t6MKIYnKAbYx8ebpFDlD4BNb0Fli0ZN+Vn6V33LglzraBMpYe3+gmZDDkqc//z7B1FS1YCnu5aLlfX8ZWNLdEGhUfGjf/hIRnNzYLNFsmjHGd76umUSOdTbjY/nDCMhyLMl/R/r8zYmwMD2E9a/LaZNGd2zF6uZ8f4uq/DFtx8czNwl2c0j1+9OlrBwRjo39mm/uUN3uBbXmYeblkHRfhDdcc9cUFq9DUsI4INHMqhvNDff/LpKj9qEa4A99TpM+GFqamrnK10LjIXwxVPw/iRYMhUWjFW6oXdEUJLSp7M1caNg/1LY+kdoqGWzdjQD1obw6Je1lFY34Fl2qNmIA/jkfM4Tg6wfJ3UaF4brcngprZrkEE8GRvny8m0p9A33ZlRiIIuzBuKva+RXt/ThVHE1b27JYc+ZUp67uTfSItG4CAI8dNzaP4wiYz3PrtzPb9Yc5kRRJTf3DSW7sBF5wy+tttmQMJFZm8y4V57jyVNzSdYU8sr3dXjo9bx5ZxxJwR5E+un53c3RTPTN5+H3d/NNTinVDWZ+PFfOW1tOUG+yjdrIK6vh5S8Oc+O8H5j07gFW7C+hutECfabAY9uV5sFztmEIT2FCO26GW/oGM7evmTCNEhNvdS5ICedse3mK/Ox2i2m1nhjOL6th3qbj3P329/zvlz9RUt1AXnmtzWf25ZZTVt1SRjivrIb3tp22Wqe20UKdyUJFXaNV2dvWWu8ZHIVB17J9Lzctt/W3Lpmbfa7MJnzxza9PMCopyGq9RTvOYOqiL/lK6LHrrBUebtorNuJgH63dwZ56HcaQr1ixomc2VJANR1tl+9WUwtevKKPz9vCPgwc/g5TbwScKBtwPkRmQswncfSGwF9tOljdPxA2K8sWVNsau3si4kiXMvyOKYXG+TEkNYuH0BBIL1uB5IZuK2kYGx/qTlRnHx3OGs2BGOsOTwvAKiWfhd6c5cl6JWjhTUsNfNh7HS6/lxIVKVuzJ5aU1h9iXW948+bQ3t5xwXz3GegvFAYO5FD4hk25hs+sYjhVW8epeV870e5IJR3/FK/GHsZgaGeBRysr49XyRuI6Hd05RfOxt2HT0AqXV1nXTpZR8vDuXJTvPYbJIymsaeW7VAfadK1e2HdALYkYo3ZyEYFRiEPdnKP5pIWBqWggTatfj+vYw+OdNcHYHq1evbtmAEEqKfBtEyhTG9wmmtUfKVSOY3F+J1jGZLby3/TTzNp2goKKOHSdLeHDBTnzbmZTrG+6NT6vlbTMD+oZ7M+uGeJ5ctpcxf/qG335+mNxS5Xxpfd72j/Ll07mZ/Glaf/58dxor5w6nb4R1CGBJlW0N7vMVdQR4WBs6vU5j4267FvTYdXYNcCatYF+9DuNa6THKztouy98D9RWg6yCmNzwN7noPSs8ooXg//hMSxsONL0JAAokhZ5pXHZkYyH5LHGO17lbG0D28L2mWo4yK6826n4w8tOwEU1MeZXaohtkfZbNk1jAGRfthtkhyLlQqzQrMFpteo1X1JnIuVHGmpAaNEDzaV8NwQznullrOaqJ5aacLEokGMFp0NEzfBNUl6OPSeeJPPwBwuKiOO7b4k5X2KuPDvAlzMxO2PgtRflYpLDX6eQr1ScBBq233i/TBu02NjtLqBlZm59GWvblljEi0bf4R6uPOb29LIWuQD7Iin5iDf8X9u6bU/7IzsPx+0kfOt/5Q4s0w5AhkL1QmcDPmQvxYBun9WTF7GOsPFuKmdeHm1FDSmnzdhcY6luy0Ptb1JgsWCzwxNoH535xESvAzuPL7O1LxNbRElkT66vmvkbG8s1XJVJ3cL6y5yw7A0p3n8NBpeGGSbcBk7zBveod1XJc6rZ0GzHcPjuKLAy3djoSArMxY1Zes0mUcxpD7+/t3vtK1ILi37bKkSaAPsF3eGq0bBCfDmF9yyHM4qQOHgU4JBRuVFERCoAcnL1Zjskh++S28M/Yjks8uRV+dz9mE+yjzv4GDRWZe39Lid1152EhSZC9cNZUcKTAS7OXGb9ccbu4d+fKUFNy0LtSbrB+xz5bWYrZI+hrKmLD/WbQlipGJcdHw7vhF5Hh44FuXz+pj9by128g703ox0sOH4fEBfHdSqUFurDPxxk4jSfHxVGrc+Dp1ARFcwNPbl+C4voS66ZmUGsrGI0Xc2j+MhCBPhsX7o2sT4WHQaYkL8mhVMEvhck0L3LQaks4uV+rA57Sp31JbRpC2zdOATzhMfA2GzgGE0h5M44oOJZkrI8722LlqXPByd7V5gjBLyc/G9mJSahjG2kai/A0ttcub0GpceGREHDEBBjYcKrSJ8gH4JDuPmTfEX/F5mxbpy/wHBvHq+qOUVTcwIzOW+zOiGdcnmC8PFVJvsnBLv1AGRF3e/3y19Nh1dg1wJq1gX72ivRTz6016errcs2dPj28XgNpy2PUubHsdzI0QkQ53zFfCCbtBfnkNP52vxN3VhZ8v20dJdQP9I7wINGg4VdrA3DEJrN1fwPacEqvPDU8IoL7RwozMGKrrTPzqs0PN78UGGJieHsUfW40GHx0Zy4G8CnqHejPLby/RX//M6vssoWmcm/BPvjt5kRc3K6nTf70njbsGRXKkoII5i7PJLa1F4yJ4YmwCk1JDuXP+DqtU8CGx/iyckY5ZSo4XVvI/Xxzh8HkjWhfB42N78ciIWGUEa7FA6UmMRWf4vkjDvH1w9EItySGevPdw+uVjmjf+RrkRbvuTdUEsF62SnNNJ04GusDI7j2c+aYm/7xXkyaJHhjS3U7scB/LKmfXhHqL9DIxODuLPG63Tr/tH+LBk5lCric8robSqnnqThRBvd3XkrdJlhBDZUsr0tssdxke+ePHintmQ3hdGPg2PfQdztimFoq7QiLenNcLXwLg+IYzoFcTCrHQGRvlyIL8SY6Ng3r0DCfVxJzXC9rE6IcgTi7SQFunLN8eLrd47U1LD6YvVfDo3k7/fO4Dls4cxPT2KA3kVrMzOwxfb+scuxlzWHixqNuKAEgcPpIT7sOqxTFbNHc66J0fyxJhe5JXV2tTz2H2mlPzyWgw6Dct2neNwk4/eZJG8sflEczw1JzbAOzfg/ck0Jn47jVXDTrHskUEs/K8hnSamkDwJjnwOQ+daL7/p96zbeW1qVkxKDWXprKE8MyGJv9ydxoKs9C4Z8XqTmfnfnKTIWM/us2WUVje0NLdA8cU/f0tvvPSuV33e+nu6Eearv2ZGvNFs4WiBka+OFLL3XBnV9e0np/TYdXYNcCatYF+9DuNaqanpYLLxeqBx7dYIvDOtaVF+fPBIBuW1jfgZXJtrP0f46PnqSBE5xUpIYHyggTFJgdw5MJwIPz0jEwPZeMS652VMoAeDYvwYFKM8apstktfu6sezKw9wlFjadq5s6P8Qaw63pFDf2j8MbWUhNCXAB3u7W7UZa68utYdOg16noaymkS3Him3ez7lQxaigGlj9mJLWDmAxY/jqWYbPGQq+XailETEYbn4Njn4BU95QfPM+URA+iPwPl3X++S7g4aYlMyGQzISuNeq+RE2DmUP5Fc2v399xhoeGxZCVGYvBTUtSsCd9mvzglzsXGk0WTFKib6fEwrXmqyNFPLlsb3NewVPjEpkzKh5Dm1j0Hr3OuokzaQX76nUYQ/7vhrfetTku+RK9QrxYMnMoBwsqOF5YSVFlHY8v3YvJbGFh1hDGJgeTmVDIjiY/9oAoX5sCQxoXwZS0cFLCvamurqF+6ge4bf41VBfD4EdwHTaTtwb6cbq4Ch+9K8mhXnyy5AOgTYW/JpJDPLmlXyjrD7Zkir44uQ/R/gbqGi0MjLJ9UojyN0B1rlWIJaDU3q4shNB+ne8grU5pvxaQBHsWKFUWkTB0LmkJNk+OPYqv3pXb08J56xtlPkNK+PD7syzMGsKNXWiRJqUk+2wZ7247RUFFLQ8Pj2Vcn2CbyJRrRW5pDc+vOmCVHPb3zScY2zvouvnaVRwLh/GRNzQ0oNNdXU2Knqa7Wtfsy+ep5fuslvUJ82L57OFIKTlVXI1FSuKDPPDvysVffVGJkPEKU0a2V6i3uLKOg/lGLlbVExfgQd8Ibww65R5/MK+CrPd3UdI0aTglLZxfT+5DsKUY3h0NNa18/lfj3z7wMXw6y2qR5bY3cRn0UAcf6BnOllTz+pc/sf5QITqNC78Yn8j9GTFWYYrQ/r49mF/OXfN3WKWmv3xbX7IyY6+L1oP5FUz5x7c2yxc8nM74FOukov+k66yn6Qm9Du8j37ZtW+crOQjd1WqsbbRZVmSsp77RjK9Bx6AYP9Jj/btmxEEpiuUT2a4Rh871Bnm5c2PvYO5Jj2JInH+zEQcl5HDNz0bw0aMZrJo7nD/cnqq4ZnyjlBKtTbXJcdXDHf/XkrbfVQ6ttlkkDtg/fjgmwIM/3z2AjU+PYuMvRjFnVIKNEYf29+3+3AorIw7w9jcn240hvxaEersR5W8dJeSqETbROPCfdZ31NPbU6zCulVOnTtlbQpfprtaUcB8eHh5DgKcbZosFnUaDm1YQ5HV9Hr27qzfSz9D+JGH8GJi9FSrPKzcT/wS40iSW8AFwfL3yOa8wqC2jLqAvHQcv9hx6nVIO9nK0t2/d2qk7YnDToLlO0SlBXu68ce9Anlq+l3OltfgZXPnj1P7tllP4T7rOehp76nUYQ/6fRJCnjt1nSjl6Xokm0boI3ns4XWmj5mz4xSh/V0vK7VRcLCA7cAo7L7oR7yPoGxmAcyVnWzMgyhdfg6tV3e7/vinJKunoWjMw2o9P546gyFiHr0FnUxtc5d8bhzHkEydOtLeELtNdrYcKjM1GHJSwvj9tPMagGD98rjIu+XJ0V6+USmNks0US4atvt9LdVX93UDLLA57gtS9PAIrroU+ohffDYztsgeZItLdvE0O8WD57GN/8VEyRsY7xKSEMjLYNPb3WBHq5EdjJU91/0nXW09hTr8MY8srKys5XchC6q7WsTbYhKIWaahtMV23ILRbJ2dJqquvNRPjp8Ws1+uuO3vKaBj7ek8e8TcdpMFl4YGg0c0YnXLbBwZWQW1rD31pluwIcLazk6PlKpzDkHe3b3qHe9A7tOFXfXvwnXWc9jT31dmtoJYT4mxDiFSHEC0KI5UKIrlXQb4cdO3Z0R0qP0l2tyaG2fte7B0cSdJUNBGoaTCzddY5Jf9/Orf/4lgcX7ORYYUuyUHf07j5Txqvrj1LTYMZkkXzw/VnWHrAte3u1mCySBpNtlb+2ZQlA+Z37c8v56kghRwoq2v1cT+NM5y04l15n0gr21dvdZ+RqKeWLUsrXgL3Ai9dA0789/SJ8+L8HBxHm446rRvDA0GiyMuOuejLsSIGRlz471JyhebjAyKvrf6K6nT6VV8q3J2wTglZl53eYOXilRPjpuWuQdR9Kb3ctSSHWE3V1jSY+2HGG29/6jlkfZnPrP77lX4fOY4/wWRUVR6NbrhUp5UutXroAVR2t2xlpaWndkdKjdFerm6uGSalhDInxp85kJsTb3aq915VytsQ2o2zr8WJKqhrw8Nd2S29CkG3kQ0q4V7uRGVeDm1bDU+MTifY38NnefFLCvblvYBDxbbZ7oqjKqgKhRcKvPj3IgEhfYgI7KQdwHXGm8xacS68zaQX76u3UkAshNgDttSr5jZTy86Z1fIEJwNTLfM9sYDZAaGgo7777LgAZGRkEBQWRnZ3N/v37iY6OZsKECSxYsAAAnU5HVlYWn332GRcuKFUBp06dSk5ODvv3KwWRMjMz8fLyYsOGDQDEx8czatQoFi1aBIDBYODBBx9k5cqVlJYqNUimT5/OoUOHOHxY6bE4evRotFotmzdvBiApKYkhQ4awZMkSALy9vbn33ntZvnw5paWl7N+/nwceeIDdu3dz/LhSG2TcuHGYTCa2bt0KKB1DUlNTm+sU+/v7M23aNBYvXtyczpuVlcXWr7c0hy5NnDiRysrK5se0tLQ0evXqxapVqwAIDg7mjjvuYNGiRTQ0KL723mPusNnfySGeHMzeyZenT5CcnExRURFr1qwBIDw8nFtvvZUFCxZgsVhwcXFh5syZrF27loICxW1y++23k5eXR7AUxPjrOdvUDd3bXcvN8e5s+mrjNTtO//pkMQEeHryQHs3okWms/HgFh7crY4Kp99zL8fyL1EpXEoI8ybnQMlaobjBzIvc8gfqwdo+T0ai4l67Vcdq2bZvNcbp03nblOM2cOZONGzdy7tw5ACZPnkxxcTG7du0CYPDgwURGRl7VccrOzgZarqd165Sqkm2vJyEEQ4cOdajrqaPjNGTIEI4dO3ZNjtOVXE9Xe5z27t3L/v37r8lx6uh66hApZbf+AB9gKRDX1c8MHjxYtuWdd96xWeaoOJrWi1V18plP9smY59bKmOfWyj6//pfceepi8/vd1ZtXWiM3HSmU6w8UyJMXKrsrt1Mu6T1XUi2fWvajjH1+rUz81Xr54qcH5OwPdzf/zoxXvpKFFbXXXU9XtDoLzqTXmbRK2TN6gT2yHZvaLdeKECIQmAc8K6XMF0JMlVKu6s53qlw5AR5uvDQ5henpURhrG4kN9LBxTXSHCD+9XeKSP9uXz2f7lBFNg9nC4p3neHZiMpuOXiDMx52/3zuAEO/OJ4hLq+s5mF9BflktUf4GEoM9yS2t5XxFLZF+BpJDvRyy0bGKSlfp7tm7sek7ljQls1QCV2XIg4M7L0bkKDiiVh+9K+mx7Re2d0S9lyM4OJiqukY+32cbHVNkrGP9z0fi56EjuAtRPtX1Jt7YfIJFO5RuQf0jfRga589721t6cr40uQ8zMmOvap7CGfets+BMWsG+eh2maJaKSmtMZgsvfHqQT9q0kXvtzn7cNzS6y99zIK+c2978rvn10+MTmbf5BK1Pe62LYP1TN5AUcvl0fBUVe+PwRbMuTaQ4A86kFZxTr1bjQtaIWPw9WhKb+oZ5k9mrk5Z8baiut26E3WiRtB27mCyy3UJmXdXqTDiTXmfSCvbV6zCOwUszxs6AM2kF59XbN9yH1Y9ncqKoCp3WheRQry75xFsTE2AgxNuNIqOS/t9osuCt12KsbYmDD/Jyu+o5AGfdt86AM2kF++p1mBG5ikp7xAR4MD4lhFFJQVdsxAHCffX8c8YQRiUFonfVUFBey1v3DyIhSIk97xPmxXsPDybMCcoBqKh0hMP4yC/FyToDzqQVVL0A1Q0mKmoa8TPo0Os0lFbXU1bTSICHrltVCdV9e/1wJq3QM3od3ke+ceNGe0voMs6kFVS9AB46LeG+evQ6pfmGv4cbCUGe3S4tq+7b64czaQX76nUYQ34pi8oZcCatoOq9njiTVnAuvc6kFeyr12EMuYqKiorK1eEwPvL8/HwiIiJ6XMvV4ExaQdV7PXEmreBcep1JK/SMXof3kRcX25ZLdVScSSuoeq8nzqQVnEuvM2kF++p1GEN+qbKYM+BMWkHVez1xJq3gXHqdSSvYV6/DGHIVFRUVlavDLj5yIUQxcLbN4kDgYo+LuTqcSSuoeq8nzqQVnEuvM2mFntEbI6UMarvQLoa8PYQQe9pz4jsizqQVVL3XE2fSCs6l15m0gn31qq4VFRUVFSdHNeQqKioqTo4jGfJ37S3gCnAmraDqvZ44k1ZwLr3OpBXsqNdhfOQqKioqKleHI43IVVRUVFSuAtWQq6ioqDg5DtMhSAjxN6AGqALSgF9IKQvtq6pjhBAuwCzg98CNUspDdpZkgxBiPHAXcAGQUsr/sbOkyyKECAX+AKRJKYfYW8/lEEIkoGj9EYgESqSUv7OvqvZpOle/AHYCOiABeERKWWtXYZ0ghNCjaN4opXzG3nouhxDiB6Cu6aVZSjmuJ7fvMIYcqJZSvgQghHgOeBF40r6SLksayklWY28h7SGEMABvA32llPVCiFVCiHFSys321nYZRgJrgAH2FtIF/IHlUso1AEKII0KIdVLKbDvr6ojvpZR/ABBCrEG5wS+xr6RO+QOw194iusiXUsqX7bVxhzHkl4x4Ey4oI3OHRUq5F0AIYW8pHTEcOCulrG96/R0wGXBYQy6lXCmEGGNvHV1BSrm7zSIXoNoeWjpDSmlBMYoIIbQoTxDH7CqqE4QQD6Gcs/0BTzvL6Qr9mgagemC3lHJdT268Rw25EGIDENLOW7+RUn7etI4vMAGY2pPa2qMreh2YYKCy1Wtj0zKVa4wQ4k5gg5TyJ3truRxCiInA08BaKeWezta3F0KIFKCPlPJXQoj+9tbTRf4opdwlhNAA24QQlVLKbT218R415FLKiZd7XwjhA8xH8d+V9oyqjulMr4NzAfBq9dq7aZnKNUQIMRYYC/zC3lo6Q0q5AdgghPhQCPG4lHK+vTV1wJ1AnRDieRR3m04I8Qsp5Tw76+oQKeWupn/NQojtKOfEv6chvxxCiEBgHvCslDJfCDFVSrnK3rqcmO+BGCGEW5N7ZQTKTVLlGiGEmAzcADwFhAkhYqSU39tZlg1NI9y4Vo/7p4F4O0q6LFLKVy79XwjhDng6shEXQvQGRkgp/9m0KBH4tCc1OIwhBzai6FnS5HeuBBzWkAsh/IAnAB9gthBiqZTyBzvLakZKWSOEmAu80VRt8oCDT3QihBgNPIRiFF8C/uKokRVCiMHACmAP8DXgAbyFcgN1NOqBR4UQAwFXoA/wc/tK6hwhxFRgFMqI/D4p5TJ7a+oAI3CrECIc5ck3F+hRrWpmp4qKioqToyYEqaioqDg5qiFXUVFRcXJUQ66ioqLi5KiGXEVFRcXJUQ25ioqKipOjGnIVFRUVJ0c15CoqKipOzv8HllFz+RUWeXwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_pca(X_dev@projs_inlp[0][1], y_dev, path=\"analysis/analysis-results\",title = \"INLP\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  FutureWarning\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_pca(X_dev@projs_rlace[0][1], y_dev, path=\"analysis/analysis-results\",title = \"RLACE (ours)\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot_pca(X_dev@projs_words[2], y_dev, path=None,title = \"PCA-based debiasing\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import matplotlib.pyplot as plt\n",
    "# plt.rcParams['font.family'] = 'Sans'\n",
    "\n",
    "# accs_inlp = np.array([projs_inlp[i][\"accs\"] for i in projs_inlp.keys()])\n",
    "# accs_rel = np.array([projs_rlace[i][\"accs\"] for i in projs_inlp.keys()])\n",
    "\n",
    "# accs_inlp_df = pd.DataFrame(accs_inlp, columns = range(len(projs_inlp[0][\"accs\"])))\n",
    "# accs_rel_df = pd.DataFrame(accs_rel, columns = [0]+projs_rlace[0][\"ranks\"])\n",
    "\n",
    "# accs_inlp = accs_inlp.mean(axis=0)\n",
    "# accs_rel = accs_rel.mean(axis=0)\n",
    "\n",
    "# ranks = projs_rlace[0][\"ranks\"]\n",
    "# ranks = [0] + ranks\n",
    "\n",
    "# ax = sn.lineplot(range(16), accs_inlp[:16], marker=\"o\", label = \"INLP\")\n",
    "# sn.lineplot(ranks, accs_rel, label=\"RLACE (ours)\", marker = \"o\", color = \"red\")\n",
    "# ax.axhline(get_maj(y), label=\"Majority\", color = \"black\", linestyle=\"--\")\n",
    "# sn.set(font_scale=1)\n",
    "# ax.set(ylabel=\"Accuracy Post-Projection\", xlabel=\"Dimensions Removed\", ylim=[0,1.03])\n",
    "# plt.xticks(range(0,21,2), fontsize=16)\n",
    "# plt.yticks(fontsize=20)\n",
    "\n",
    "# plt.xlabel(\"Dimensions Removed\", fontsize=20)\n",
    "# plt.ylabel(\"Accuracy Post-Projection\", fontsize=20)\n",
    "\n",
    "# plt.subplots_adjust(bottom=0.18)\n",
    "# plt.subplots_adjust(left=0.18)\n",
    "# ax.legend(fontsize=15)\n",
    "# ax.figure.savefig(\"relaxed_vs_inlp_{}.pdf\".format(\"glove\"), dpi = 400)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "# accs_rel_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original 2: 1.0\n",
      "Original 3: 0.6778304418036202\n",
      "Original 5: 0.6321732227062642\n",
      "Original 10: 0.45781218541489044\n",
      "Original 15: 0.38969314341437383\n",
      "Original 20: 0.3656262343542788\n",
      "Original 30: 0.33277933587229375\n",
      "pca 2 0.025264582731482788\n",
      "pca 3 0.030678681888024067\n",
      "pca 5 0.027258495517400706\n",
      "pca 10 0.025463666223980858\n",
      "pca 15 0.02301553471617879\n",
      "pca 20 0.021677363169225083\n",
      "pca 30 0.019969663067186692\n",
      "Starting\n",
      "2 5.5354631229660656e-05\n",
      "3 0.00037905066629563105\n",
      "5 0.03586216148073483\n",
      "10 0.07369829513154451\n",
      "15 0.10654785147777956\n",
      "20 0.11748296468127274\n",
      "30 0.1089437326328201\n",
      "-------------\n",
      "2 0.0012751659333405175\n",
      "3 0.0011013075648533765\n",
      "5 0.0619715413385864\n",
      "10 0.12817636827930912\n",
      "15 0.12234756768209662\n",
      "20 0.14284760974657137\n",
      "30 0.13502077547759936\n",
      "-------------\n",
      "==========================\n",
      "2 7.695006110047016e-05\n",
      "3 0.0004182032666343437\n",
      "5 0.05336971873022601\n",
      "10 0.07572290171654525\n",
      "15 0.09255854272649536\n",
      "20 0.11354575812991977\n",
      "30 0.11081369170892007\n",
      "-------------\n",
      "2 0.0012264177477130161\n",
      "3 0.0014788382095883837\n",
      "5 0.07155968424828452\n",
      "10 0.16854270577378372\n",
      "15 0.1714921411683711\n",
      "20 0.1484756406442243\n",
      "30 0.14921041269033014\n",
      "-------------\n",
      "==========================\n",
      "2 5.5354631229660656e-05\n",
      "3 0.00035751846848166094\n",
      "5 0.03472300088096338\n",
      "10 0.10616236499043373\n",
      "15 0.10819024109461722\n",
      "20 0.10265156693796593\n",
      "30 0.11044716446147597\n",
      "-------------\n",
      "2 0.0011845422071501109\n",
      "3 0.0016149726638388902\n",
      "5 0.1175777325474793\n",
      "10 0.18074696938448484\n",
      "15 0.15311471383613276\n",
      "20 0.15984074019035988\n",
      "30 0.14825557238846798\n",
      "-------------\n",
      "==========================\n",
      "2 7.695006110047016e-05\n",
      "3 0.0004784621554715247\n",
      "5 0.052225816609496684\n",
      "10 0.0838206211651962\n",
      "15 0.13529675133293512\n",
      "20 0.11114830824854975\n",
      "30 0.11252025037365496\n",
      "-------------\n",
      "2 0.0012751659333405175\n",
      "3 0.0014279529489659016\n",
      "5 0.1089630877913866\n",
      "10 0.12630311173929473\n",
      "15 0.13704353409742964\n",
      "20 0.13734696978475752\n",
      "30 0.13064207135972503\n",
      "-------------\n",
      "==========================\n",
      "2 3.73943741131813e-05\n",
      "3 0.00048629035268796575\n",
      "5 0.0244175096688183\n",
      "10 0.07481337196777003\n",
      "15 0.13333132622888258\n",
      "20 0.11321275180592726\n",
      "30 0.11103481038874169\n",
      "-------------\n",
      "2 0.001180620595266846\n",
      "3 0.0009518509984958149\n",
      "5 0.07845414988831904\n",
      "10 0.1351094961503922\n",
      "15 0.13754725883473784\n",
      "20 0.14133490073176355\n",
      "30 0.14578284714833037\n",
      "-------------\n",
      "==========================\n"
     ]
    }
   ],
   "source": [
    "\n",
    "v_measures_rlace = defaultdict(dict)\n",
    "v_measures_inlp = defaultdict(dict)\n",
    "v_measures_pca = defaultdict(dict)\n",
    "v_measures_original = defaultdict(dict)\n",
    "\n",
    "number_of_clusts = [2,3,5,10,15,20,30]\n",
    "\n",
    "NUM_CLUSTERS=5\n",
    "\n",
    "for num_clust in number_of_clusts:\n",
    "    set_seeds(0)\n",
    "    after = get_v_measure(X_dev, y_dev, np.eye(300), num_clusts=num_clust)\n",
    "    v_measures_original[1][num_clust] = after\n",
    "    print(\"Original {}: {}\".format(num_clust, after))\n",
    "\n",
    "for num_clust in number_of_clusts:\n",
    "        set_seeds(0)\n",
    "        after = get_v_measure(X_dev, y_dev, projs_words[0], num_clusts=num_clust)\n",
    "        v_measures_pca[1][num_clust] = after\n",
    "        print(\"pca\", num_clust, after)\n",
    "        \n",
    "        \n",
    "print(\"Starting\")\n",
    "rank = 1\n",
    "for rand_seed in projs_rlace.keys():\n",
    "    \n",
    "    set_seeds(0)\n",
    "    Ps_rlace = projs_rlace[rand_seed]\n",
    "    Ps_inlp = projs_inlp[rand_seed]\n",
    "    \n",
    "    P_rlace = Ps_rlace[rank]\n",
    "    P_inlp = Ps_inlp[rank-1]\n",
    "    \n",
    "    for num_clust in number_of_clusts:\n",
    "\n",
    "        after = get_v_measure(X_dev, y_dev, P_rlace, num_clusts=num_clust)\n",
    "        v_measures_rlace[rand_seed][num_clust] = after\n",
    "        print(num_clust, after)\n",
    "    print(\"-------------\")\n",
    "    \n",
    "    #for num_clusts in number_of_clusts:\n",
    "    for num_clust in number_of_clusts:\n",
    "        set_seeds(0)\n",
    "        after = get_v_measure(X_dev, y_dev, P_inlp, num_clusts=num_clust)\n",
    "        v_measures_inlp[rand_seed][num_clust] = after\n",
    "        print(num_clust, after)\n",
    "        \n",
    "    print(\"-------------\")\n",
    "    \n",
    "\n",
    "        \n",
    "    print(\"==========================\")\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "v_measures_inlp_df = pd.DataFrame(v_measures_inlp)\n",
    "v_measures_inlp_df['average'] = v_measures_inlp_df.mean(numeric_only=True, axis=1)\n",
    "v_measures_inlp_df[\"std\"] = v_measures_inlp_df.std(numeric_only=True, axis=1)\n",
    "v_measures_inlp_df = v_measures_inlp_df.rename_axis(\"num_clusters\").reset_index()\n",
    "\n",
    "\n",
    "v_measures_rlace_df = pd.DataFrame(v_measures_rlace)\n",
    "v_measures_rlace_df['average'] = v_measures_rlace_df.mean(numeric_only=True, axis=1)\n",
    "v_measures_rlace_df[\"std\"] = v_measures_rlace_df.std(numeric_only=True, axis=1)\n",
    "v_measures_rlace_df = v_measures_rlace_df.rename_axis(\"num_clusters\").reset_index()\n",
    "\n",
    "v_measures_pca_df = pd.DataFrame(v_measures_pca)\n",
    "v_measures_pca_df['average'] = v_measures_pca_df.mean(numeric_only=True, axis=1)\n",
    "v_measures_pca_df[\"std\"] = v_measures_pca_df.std(numeric_only=True, axis=1)\n",
    "v_measures_pca_df = v_measures_pca_df.rename_axis(\"num_clusters\").reset_index()\n",
    "\n",
    "\n",
    "v_measures_original_df = pd.DataFrame(v_measures_original)\n",
    "v_measures_original_df['average'] = v_measures_original_df.mean(numeric_only=True, axis=1)\n",
    "v_measures_original_df[\"std\"] = v_measures_original_df.std(numeric_only=True, axis=1)\n",
    "v_measures_original_df = v_measures_original_df.rename_axis(\"num_clusters\").reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "v_measures_inlp_df.to_csv(\"analysis/analysis-results/v_measure_inlp.csv\")\n",
    "v_measures_rlace_df.to_csv(\"analysis/analysis-results/v_measure_rlace.csv\")\n",
    "v_measures_pca_df.to_csv(\"analysis/analysis-results/v_mesure_pca.csv\")\n",
    "v_measures_original_df.to_csv(\"analysis/analysis-results/v_mesure_original.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_clusters</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>0.000055</td>\n",
       "      <td>0.000077</td>\n",
       "      <td>0.000055</td>\n",
       "      <td>0.000077</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.000060</td>\n",
       "      <td>0.000015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>0.000379</td>\n",
       "      <td>0.000418</td>\n",
       "      <td>0.000358</td>\n",
       "      <td>0.000478</td>\n",
       "      <td>0.000486</td>\n",
       "      <td>0.000424</td>\n",
       "      <td>0.000052</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>0.035862</td>\n",
       "      <td>0.053370</td>\n",
       "      <td>0.034723</td>\n",
       "      <td>0.052226</td>\n",
       "      <td>0.024418</td>\n",
       "      <td>0.040120</td>\n",
       "      <td>0.011099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>0.073698</td>\n",
       "      <td>0.075723</td>\n",
       "      <td>0.106162</td>\n",
       "      <td>0.083821</td>\n",
       "      <td>0.074813</td>\n",
       "      <td>0.082844</td>\n",
       "      <td>0.012195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>0.106548</td>\n",
       "      <td>0.092559</td>\n",
       "      <td>0.108190</td>\n",
       "      <td>0.135297</td>\n",
       "      <td>0.133331</td>\n",
       "      <td>0.115185</td>\n",
       "      <td>0.016548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>20</td>\n",
       "      <td>0.117483</td>\n",
       "      <td>0.113546</td>\n",
       "      <td>0.102652</td>\n",
       "      <td>0.111148</td>\n",
       "      <td>0.113213</td>\n",
       "      <td>0.111608</td>\n",
       "      <td>0.004925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>30</td>\n",
       "      <td>0.108944</td>\n",
       "      <td>0.110814</td>\n",
       "      <td>0.110447</td>\n",
       "      <td>0.112520</td>\n",
       "      <td>0.111035</td>\n",
       "      <td>0.110752</td>\n",
       "      <td>0.001147</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   num_clusters         0         1         2         3         4   average  \\\n",
       "0             2  0.000055  0.000077  0.000055  0.000077  0.000037  0.000060   \n",
       "1             3  0.000379  0.000418  0.000358  0.000478  0.000486  0.000424   \n",
       "2             5  0.035862  0.053370  0.034723  0.052226  0.024418  0.040120   \n",
       "3            10  0.073698  0.075723  0.106162  0.083821  0.074813  0.082844   \n",
       "4            15  0.106548  0.092559  0.108190  0.135297  0.133331  0.115185   \n",
       "5            20  0.117483  0.113546  0.102652  0.111148  0.113213  0.111608   \n",
       "6            30  0.108944  0.110814  0.110447  0.112520  0.111035  0.110752   \n",
       "\n",
       "        std  \n",
       "0  0.000015  \n",
       "1  0.000052  \n",
       "2  0.011099  \n",
       "3  0.012195  \n",
       "4  0.016548  \n",
       "5  0.004925  \n",
       "6  0.001147  "
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "v_measures_rlace_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_clusters</th>\n",
       "      <th>1</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>0.677830</td>\n",
       "      <td>0.677830</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>0.632173</td>\n",
       "      <td>0.632173</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>0.457812</td>\n",
       "      <td>0.457812</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>0.389693</td>\n",
       "      <td>0.389693</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>20</td>\n",
       "      <td>0.365626</td>\n",
       "      <td>0.365626</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>30</td>\n",
       "      <td>0.332779</td>\n",
       "      <td>0.332779</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   num_clusters         1   average  std\n",
       "0             2  1.000000  1.000000  0.0\n",
       "1             3  0.677830  0.677830  0.0\n",
       "2             5  0.632173  0.632173  0.0\n",
       "3            10  0.457812  0.457812  0.0\n",
       "4            15  0.389693  0.389693  0.0\n",
       "5            20  0.365626  0.365626  0.0\n",
       "6            30  0.332779  0.332779  0.0"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "v_measures_original_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_clusters</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>0.001275</td>\n",
       "      <td>0.001226</td>\n",
       "      <td>0.001185</td>\n",
       "      <td>0.001275</td>\n",
       "      <td>0.001181</td>\n",
       "      <td>0.001228</td>\n",
       "      <td>0.000041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>0.001101</td>\n",
       "      <td>0.001479</td>\n",
       "      <td>0.001615</td>\n",
       "      <td>0.001428</td>\n",
       "      <td>0.000952</td>\n",
       "      <td>0.001315</td>\n",
       "      <td>0.000248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>0.061972</td>\n",
       "      <td>0.071560</td>\n",
       "      <td>0.117578</td>\n",
       "      <td>0.108963</td>\n",
       "      <td>0.078454</td>\n",
       "      <td>0.087705</td>\n",
       "      <td>0.021692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>0.128176</td>\n",
       "      <td>0.168543</td>\n",
       "      <td>0.180747</td>\n",
       "      <td>0.126303</td>\n",
       "      <td>0.135109</td>\n",
       "      <td>0.147776</td>\n",
       "      <td>0.022468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>0.122348</td>\n",
       "      <td>0.171492</td>\n",
       "      <td>0.153115</td>\n",
       "      <td>0.137044</td>\n",
       "      <td>0.137547</td>\n",
       "      <td>0.144309</td>\n",
       "      <td>0.016717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>20</td>\n",
       "      <td>0.142848</td>\n",
       "      <td>0.148476</td>\n",
       "      <td>0.159841</td>\n",
       "      <td>0.137347</td>\n",
       "      <td>0.141335</td>\n",
       "      <td>0.145969</td>\n",
       "      <td>0.007801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>30</td>\n",
       "      <td>0.135021</td>\n",
       "      <td>0.149210</td>\n",
       "      <td>0.148256</td>\n",
       "      <td>0.130642</td>\n",
       "      <td>0.145783</td>\n",
       "      <td>0.141782</td>\n",
       "      <td>0.007522</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   num_clusters         0         1         2         3         4   average  \\\n",
       "0             2  0.001275  0.001226  0.001185  0.001275  0.001181  0.001228   \n",
       "1             3  0.001101  0.001479  0.001615  0.001428  0.000952  0.001315   \n",
       "2             5  0.061972  0.071560  0.117578  0.108963  0.078454  0.087705   \n",
       "3            10  0.128176  0.168543  0.180747  0.126303  0.135109  0.147776   \n",
       "4            15  0.122348  0.171492  0.153115  0.137044  0.137547  0.144309   \n",
       "5            20  0.142848  0.148476  0.159841  0.137347  0.141335  0.145969   \n",
       "6            30  0.135021  0.149210  0.148256  0.130642  0.145783  0.141782   \n",
       "\n",
       "        std  \n",
       "0  0.000041  \n",
       "1  0.000248  \n",
       "2  0.021692  \n",
       "3  0.022468  \n",
       "4  0.016717  \n",
       "5  0.007801  \n",
       "6  0.007522  "
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "v_measures_inlp_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sn.set()\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "plt.rcParams['font.family'] = 'Sans'\n",
    "\n",
    "v_measures_rlace_df.plot('num_clusters', 'average', yerr='std', ax=ax, label=\"RLACE (ours)\", marker=\"*\")\n",
    "v_measures_inlp_df.plot('num_clusters', 'average', yerr='std', ax=ax, label=\"INLP\", marker=\"*\")\n",
    "v_measures_pca_df.plot('num_clusters', 'average', yerr='std', ax=ax, label=\"PCA-based\", marker=\"*\")\n",
    "#v_measures_original_df.plot('num_clusters', 'average', yerr='std', ax=ax, label=\"Original space\", marker=\"*\", linestyle = \"--\")\n",
    "\n",
    "plt.ylabel(\"V-measure\", fontsize=18)\n",
    "plt.xlabel(\"Number of clusters\", fontsize=18)\n",
    "\n",
    "plt.legend(fontsize=18)\n",
    "#ax.yaxis.grid(color='gray', linestyle=\"-\")\n",
    "#ax.xaxis.grid(color='gray', linestyle='-')\n",
    "plt.yticks(fontsize=18)\n",
    "plt.xticks([2,5,10,15,20,25,30], fontsize=18)\n",
    "plt.subplots_adjust(bottom=0.17)\n",
    "plt.subplots_adjust(left=0.17)\n",
    "ax.figure.savefig(\"analysis/analysis-results/v-mesaure.pdf\", dpi = 400)    \n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "rlace 1 0.5246666666666666\n",
      "inlp 1 0.862\n",
      "rlace 2 0.5255555555555556\n",
      "inlp 2 0.7617777777777778\n",
      "rlace 4 0.5177777777777778\n",
      "inlp 4 0.6484444444444445\n",
      "rlace 8 0.5008888888888889\n",
      "inlp 8 0.6122222222222222\n",
      "rlace 12 0.5048888888888889\n",
      "inlp 12 0.5762222222222222\n",
      "rlace 16 0.512\n",
      "inlp 16 0.5677777777777778\n",
      "rlace 20 0.5308888888888889\n",
      "inlp 20 0.5702222222222222\n",
      "==========================\n",
      "1\n",
      "rlace 1 0.49733333333333335\n",
      "inlp 1 0.9288888888888889\n",
      "rlace 2 0.5582222222222222\n",
      "inlp 2 0.8315555555555556\n",
      "rlace 4 0.5277777777777778\n",
      "inlp 4 0.7024444444444444\n",
      "rlace 8 0.48933333333333334\n",
      "inlp 8 0.5917777777777777\n",
      "rlace 12 0.5233333333333333\n",
      "inlp 12 0.5568888888888889\n",
      "rlace 16 0.5228888888888888\n",
      "inlp 16 0.5562222222222222\n",
      "rlace 20 0.4955555555555556\n",
      "inlp 20 0.5584444444444444\n",
      "==========================\n",
      "2\n",
      "rlace 1 0.482\n",
      "inlp 1 0.9337777777777778\n",
      "rlace 2 0.5231111111111111\n",
      "inlp 2 0.8417777777777777\n",
      "rlace 4 0.5168888888888888\n",
      "inlp 4 0.7126666666666667\n",
      "rlace 8 0.5048888888888889\n",
      "inlp 8 0.6184444444444445\n",
      "rlace 12 0.5337777777777778\n",
      "inlp 12 0.5804444444444444\n",
      "rlace 16 0.5335555555555556\n",
      "inlp 16 0.5653333333333334\n",
      "rlace 20 0.5286666666666666\n",
      "inlp 20 0.5524444444444444\n",
      "==========================\n",
      "3\n",
      "rlace 1 0.5231111111111111\n",
      "inlp 1 0.8408888888888889\n",
      "rlace 2 0.5235555555555556\n",
      "inlp 2 0.752\n",
      "rlace 4 0.48533333333333334\n",
      "inlp 4 0.6633333333333333\n",
      "rlace 8 0.5253333333333333\n",
      "inlp 8 0.5851111111111111\n",
      "rlace 12 0.5348888888888889\n",
      "inlp 12 0.5606666666666666\n",
      "rlace 16 0.5208888888888888\n",
      "inlp 16 0.53\n",
      "rlace 20 0.5064444444444445\n",
      "inlp 20 0.5395555555555556\n",
      "==========================\n",
      "4\n",
      "rlace 1 0.5264444444444445\n",
      "inlp 1 0.8746666666666667\n",
      "rlace 2 0.5228888888888888\n",
      "inlp 2 0.7684444444444445\n",
      "rlace 4 0.514\n",
      "inlp 4 0.67\n",
      "rlace 8 0.4988888888888889\n",
      "inlp 8 0.5935555555555555\n",
      "rlace 12 0.5264444444444445\n",
      "inlp 12 0.5704444444444444\n",
      "rlace 16 0.51\n",
      "inlp 16 0.5435555555555556\n",
      "rlace 20 0.5237777777777778\n",
      "inlp 20 0.5342222222222223\n",
      "==========================\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "accs_inlp = defaultdict(dict)\n",
    "accs_rlace = defaultdict(dict)\n",
    "accs_pca = defaultdict(dict)\n",
    "\n",
    "model_class = SGDClassifier\n",
    "model_params = {\"max_iter\": 150000, \"tol\": 1e-4, \"n_iter_no_change\": 50, \"alpha\": 0.5*1e-4}\n",
    "\n",
    "for rand_seed in projs_rlace.keys():\n",
    "    print(rand_seed)\n",
    "    \n",
    "    set_seeds(0)\n",
    "    Ps_rlace = projs_rlace[rand_seed]\n",
    "    Ps_inlp = projs_inlp[rand_seed]\n",
    "    \n",
    "    for i,rank in enumerate(projs_rlace[rand_seed].keys()):\n",
    "        \n",
    "        P = projs_rlace[rand_seed][rank]\n",
    "        \n",
    "        clf = model_class(**model_params)\n",
    "        X_train_clean, X_test_clean = X@P, X_test@P\n",
    "        clf.fit(X_train_clean, y)\n",
    "        score = clf.score(X_test_clean, y_test)\n",
    "        accs_rlace[rand_seed][rank] = score\n",
    "        print(\"rlace\", rank, score)\n",
    "        \n",
    "        P = (projs_inlp[rand_seed][rank-1])\n",
    "        #mlp = sklearn.neural_network.MLPClassifier()\n",
    "        clf = model_class(**model_params)\n",
    "        X_train_clean, X_test_clean = X@P, X_test@P\n",
    "        clf.fit(X_train_clean, y)\n",
    "        score = clf.score(X_test_clean, y_test)\n",
    "        accs_inlp[rand_seed][rank] = score\n",
    "        print(\"inlp\", rank, score)\n",
    "        \n",
    "    print(\"==========================\")\n",
    "    \n",
    "         \n",
    "    \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pca 1 0.5615555555555556\n",
      "pca 2 0.5653333333333334\n",
      "pca 3 0.9917777777777778\n",
      "pca 4 0.9924444444444445\n",
      "pca 5 0.9933333333333333\n",
      "pca 6 0.9926666666666667\n",
      "pca 7 0.9971111111111111\n",
      "pca 8 0.9971111111111111\n",
      "pca 9 0.9968888888888889\n",
      "pca 10 0.9968888888888889\n"
     ]
    }
   ],
   "source": [
    "model_class = SGDClassifier\n",
    "model_params = {\"max_iter\": 1800000, \"tol\": 1e-4, \"n_iter_no_change\": 50, \"alpha\": 0.5*1e-4}\n",
    "\n",
    "for i in range(len(projs_words)):\n",
    "        P = projs_words[i]\n",
    "        clf = model_class(**model_params)\n",
    "        X_train_clean, X_test_clean = X@P, X_test@P\n",
    "        clf.fit(X_train_clean, y)\n",
    "        score = clf.score(X_test_clean, y_test)\n",
    "        accs_pca[0][i+1] = score\n",
    "        print(\"pca\", i+1, score)          "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   rank         0         1         2         3         4   average       std\n",
      "0     1  0.524667  0.497333  0.482000  0.523111  0.526444  0.510711  0.017885\n",
      "1     2  0.525556  0.558222  0.523111  0.523556  0.522889  0.530667  0.013810\n",
      "2     4  0.517778  0.527778  0.516889  0.485333  0.514000  0.512356  0.014288\n",
      "3     8  0.500889  0.489333  0.504889  0.525333  0.498889  0.503867  0.011889\n",
      "4    12  0.504889  0.523333  0.533778  0.534889  0.526444  0.524667  0.010804\n",
      "5    16  0.512000  0.522889  0.533556  0.520889  0.510000  0.519867  0.008447\n",
      "6    20  0.530889  0.495556  0.528667  0.506444  0.523778  0.517067  0.013757\n",
      "============================\n",
      "   rank         0         1         2         3         4   average       std\n",
      "0     1  0.862000  0.928889  0.933778  0.840889  0.874667  0.888044  0.036988\n",
      "1     2  0.761778  0.831556  0.841778  0.752000  0.768444  0.791111  0.037701\n",
      "2     4  0.648444  0.702444  0.712667  0.663333  0.670000  0.679378  0.024259\n",
      "3     8  0.612222  0.591778  0.618444  0.585111  0.593556  0.600222  0.012807\n",
      "4    12  0.576222  0.556889  0.580444  0.560667  0.570444  0.568933  0.008959\n",
      "5    16  0.567778  0.556222  0.565333  0.530000  0.543556  0.552578  0.014129\n",
      "6    20  0.570222  0.558444  0.552444  0.539556  0.534222  0.550978  0.012957\n",
      "============================\n",
      "   rank         0   average  std\n",
      "0     1  0.561556  0.561556  0.0\n",
      "1     2  0.565333  0.565333  0.0\n",
      "2     3  0.991778  0.991778  0.0\n",
      "3     4  0.992444  0.992444  0.0\n",
      "4     5  0.993333  0.993333  0.0\n",
      "5     6  0.992667  0.992667  0.0\n",
      "6     7  0.997111  0.997111  0.0\n",
      "7     8  0.997111  0.997111  0.0\n",
      "8     9  0.996889  0.996889  0.0\n",
      "9    10  0.996889  0.996889  0.0\n",
      "============================\n"
     ]
    }
   ],
   "source": [
    "#df_originals = pd.DataFrame(originals)\n",
    "df_rlace = pd.DataFrame(accs_rlace)\n",
    "df_inlp = pd.DataFrame(accs_inlp)\n",
    "df_pca = pd.DataFrame(accs_pca)\n",
    "\n",
    "for df in [df_rlace, df_inlp, df_pca]:\n",
    "    \n",
    "    df['average'] = df.mean(numeric_only=True, axis=1)\n",
    "    df[\"std\"] = df.std(numeric_only=True, axis=1)\n",
    "    df.rename_axis(\"rank\", inplace=True)\n",
    "    df.reset_index(inplace=True)\n",
    "    print(df)\n",
    "    print(\"============================\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   rank         0         1         2         3         4   average       std\n",
      "0     1  0.524667  0.497333  0.482000  0.523111  0.526444  0.510711  0.017885\n",
      "1     2  0.525556  0.558222  0.523111  0.523556  0.522889  0.530667  0.013810\n",
      "2     4  0.517778  0.527778  0.516889  0.485333  0.514000  0.512356  0.014288\n",
      "3     8  0.500889  0.489333  0.504889  0.525333  0.498889  0.503867  0.011889\n",
      "4    12  0.504889  0.523333  0.533778  0.534889  0.526444  0.524667  0.010804\n",
      "5    16  0.512000  0.522889  0.533556  0.520889  0.510000  0.519867  0.008447\n",
      "6    20  0.530889  0.495556  0.528667  0.506444  0.523778  0.517067  0.013757\n",
      "================\n",
      "   rank         0         1         2         3         4   average       std\n",
      "0     1  0.862000  0.928889  0.933778  0.840889  0.874667  0.888044  0.036988\n",
      "1     2  0.761778  0.831556  0.841778  0.752000  0.768444  0.791111  0.037701\n",
      "2     4  0.648444  0.702444  0.712667  0.663333  0.670000  0.679378  0.024259\n",
      "3     8  0.612222  0.591778  0.618444  0.585111  0.593556  0.600222  0.012807\n",
      "4    12  0.576222  0.556889  0.580444  0.560667  0.570444  0.568933  0.008959\n",
      "5    16  0.567778  0.556222  0.565333  0.530000  0.543556  0.552578  0.014129\n",
      "6    20  0.570222  0.558444  0.552444  0.539556  0.534222  0.550978  0.012957\n",
      "================\n",
      "   rank         0   average  std\n",
      "0     1  0.561556  0.561556  0.0\n",
      "1     2  0.565333  0.565333  0.0\n",
      "2     3  0.991778  0.991778  0.0\n",
      "3     4  0.992444  0.992444  0.0\n",
      "4     5  0.993333  0.993333  0.0\n",
      "5     6  0.992667  0.992667  0.0\n",
      "6     7  0.997111  0.997111  0.0\n",
      "7     8  0.997111  0.997111  0.0\n",
      "8     9  0.996889  0.996889  0.0\n",
      "9    10  0.996889  0.996889  0.0\n"
     ]
    }
   ],
   "source": [
    "print(df_rlace)\n",
    "print(\"================\")\n",
    "print(df_inlp)\n",
    "print(\"================\")\n",
    "print(df_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEYCAYAAABLOxEiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeVhUZfvA8e+ZlU0BEVQQJTVxwSVT0VzQzLXc0zQ1y9LMDNt8KyvLNDUrcyuzfuaCppb7npqaSy65i4pb7iggsi8zw8z5/YGMkixnYGBEns91vdcL55znPPeMNPecZ5VkWZYRBEEQBIVUjg5AEARBKFlE4hAEQRBsIhKHIAiCYBOROARBEASbiMQhCIIg2EQkDkEQBMEmInEIgiAINlGcONasWVOUcQiCIAglhOLE8eWXX3Ljxo2ijEUQBEEoASSlM8fr1auHu7s7NWrUoHfv3nTs2BGdTlfU8QmCIAgPGcVPHO7u7uzcuZMBAwawYcMGQkJCGDduHOHh4UUZnyAIgvCQUfzEcefOHcqVK2f9PSYmhlWrVrFq1Sp0Oh3PP/88Xbt2xcPDo8iCFQRBEBxPceLIzZkzZ/j88885ceIEWq2Wp59+mt69e9OqVSt7xSgIgiA8RBQnjh9++IERI0YAkJSUxNq1a1mxYgVnzpxBlmXq169P79690Wg0rFq1ivj4eKZMmUKdOnWK9AUIgiAIxUtx4mjZsiVTpkxhxYoVbNu2DYPBgJeXF926deP555+nevXq2a7fv38/X375JevWrSuSwAVBEATHUJw4atWqhSRJqNVqWrVqRe/evWnTpg0ajSbH600mEy1btuTAgQN2DVgQBEFwrJw/9XOg0+kIDQ2lR48elC9fPs9rk5OTef7553Fzcyt0gIIgCMLDRfETx4gRI/jhhx8U3VSWZSIjI9Hr9fkmGUEQBKFkUZw4ZFkmJSUFABcXF1SqzCkgBoOBW7duUbVq1aKLUhAEQXhoKG6q+uOPP3j77beRJIkFCxbQtGlTABISEujYsSM9evRg/PjxaLXaIgu2OMTFpWCx2D5C2cvLjdjY5CKISNQt6hZ1i7qLt26VSsLT0zXX84oTx9q1a2nVqhUTJkygQoUK1uM+Pj5s3ryZTz/9lDlz5jBy5EjFwc2ZM4dTp05x6tQprl+/jp+fH9u3b1dcPstff/3F7NmziYiIQKfT0axZM0aPHo2/v7/N97JY5AIljqyyjiLqFnWLukXdxUXxkiNnz57lq6++ypY0sgQEBPDtt9/aPPR26tSpHDhwgCpVquDu7m5T2Sxbtmzh9ddfJz09ndGjR/Pqq69y6NAh+vfvT1RUVIHuKQiCIORO8ROHwWDItuTIf/n4+Fj7QJTatm2b9angueeeIzU11abyJpOJ8ePHU6lSJRYvXoyra+ajVevWrenVqxezZs1i/PjxNt1TEARByJvixKHX67l8+TIBAQE5nr98+bLNq+UWpCnpfv/88w/R0dGEhoZakwZA7dq1adq0KRs3bmTs2LElvt9FEB42aWkpJCfHYzZnFFud0dEqLBZLsdVXOuqW0Omc8PT0RpIkxaUUJ44OHTrw+uuv8/HHH9O8eXPrh7HRaGTnzp189913dOrUyfa4C+HkyZMAPPHEEw+ca9iwIfv37+fy5cs8/vjjxRqXUPQSDInM2v4zg2r2w11fpkDlfwlfzJCggTaXd2TdD4O0tBSSkuLw8PBGq9XZ9IFTGBqNiowMx3x4P6p1y7KF+PjbJCcnUKaM8gVqFfdxvPnmm7i4uPD666/zxBNP0LJlS4KDg2nQoAGjRo3C2dmZN998s0DBF1R0dDRAjv0uPj4+AKKf4xG16dI2ImIusOnS1gKXv5hwuUDlHVn3wyA5OR4PD290On2xJQ2haEiSijJlPElLs23UluInDjc3N5YuXcrcuXPZtGkTV69eBeDxxx+nc+fODBkyBL1eb1vUhZSWlgaQYxNZVizp6ek23dPLq+Cz3b29i//bY1xaAp9tn8o7zV/Fw9n2AQZxaQlM2ze3QOWLq+70DAN30uK5kxrPl3/NxCybred2R+5nd+R+JEmiZZUm+da55+o/3D91Kau8SpIICWiOBCBJqJBAkpAASZKQkNh6cTcW2ZJDWRWdH2+bb92bzu/IsbxWpWFxn5n5lr+fI/7WssiyBWdnJ4ckDY1G8XddUbdCarUOkG36m1KcOCDzw3jEiBHWVXIdzdnZGchsLvsvg8EAgJOTk033jI1NLtDQN2/vMsTEJNlcrrCWRqwmIuYCYYdW069Wr0KXl2UZs2zGLFswWzIwyxYy7v6/WTZjtpgxy2YyLGa2XNlBxO3zfP93GB0C2qKSVKglNSpJdfdn1X0/q+/7PfPn385l1v3DvsUEV2xEvCGRBEMC8YZE4g0JxBszf0/LyDv561Q63LQunIm6kO/r9dS5k2xKxWi59zejVWlx0ThzNPLU3aQiI5M56TXzp8yf9SodRosR830f/ipUaCUNf17ck2/dWkmDSc7AgsVab0PvIHrWeM6mvx1H/a1l1W2xWDCbM98nW321+AgAHwxoZHPZR6256OzZCKpWDcj3M8oedZ85c4rq1R/PtR/aYrFk+5tSqaQ8v0TblDjyM3PmTN566y173jJP9zdH/Xd13ryasR4Fo3aOIcNyr2PS+s0bicc9qmGWLVis/7ubCO7+bJEtxKbHZbtfVvmCOH77FMdvnyrwazkWc5JjMZn9VSpJRVldGdz1Zang4k2gZ3U8dO6468vioXfHQ1+WrVf/Yv/NQ2hUGjIsGQRXbGRT0lwSsZK9kQdQq9SYLWaaVXxScfmssll1t/BtWuC6MywZOKn1JbKf42Fy4sQxfv55NpcvX6J16zZYLBYiI2/QrVtP2rXrQFxcHDNmfMvWrZvp2rUH7733YY6Ls547F8HQoYNZvnwd3t4+2c5duHCesLD5qNUa1Go1UVG3aNDgCQYMGMzp0+HZ6s9y5MghlixZmWPMu3bt5Nixw4SGvmfX9yI3Zcu688knHzBhwld22fK7QInjzp07DzQBybLMkiVLijVx1KtXD4CjR4/y1FNPZTt37Ngx3Nzcch0FVtJ90fxDVl5Yz5GoE1iwIAEuGhfKOZXDLFtQSyq0Kg0qVdY3f7X1KUAtqQmwZHAl6Tp30uKwYEElqfBxKU9dz0BcdC6ZTwgqNRpJnfkUkfWzSo0hw8CBW4e5lHCFDNmMRlLzmHtVmlVsjJNGny1p3Z+ssp5aUk2phN8+w83UaCyyBbWkJtCzOt2qd8bPrRIqKe/H8lRTKi39mtG17tOsO7WdREOiTe9dkjGJln7NaOkbzJ7IAzaVzyrriLofBRlmC5GxKSQkG3B3s0/Tdv36Denc+TlWr17B6NFjAIiMvMHAgX1p1aoNnp6eDBs2gq1bN/P226NzXdF748Z1+PtXYePGdQwe/Kr1+LVrVxkz5n9Mnfo9FStWBODOnVheeeVFevXqm2P9AD/9lPPafmlpacyc+R1hYcvs8vqV8POrzJNPNiYsbB6vvvp6oe9nU+KYO3cuv/zyC3fu3Cl0xbaKjo4mKSkJX19faxNVkyZN8Pb2Zvny5bz88svWIbkREREcPHiQXr16PbJDcd31ZXFSO1mTBkg08qlv+7fftMxvz2aLmcfdq9GrZldFZS8nXuNC/CW0Ki0ZlgwquvjQzLex4rpTTGlEpkRZ6/ZyKod/GT9FZYfVHwyAt2cZ+gX2VFznf8sDNpd3ZN2PgtsJ6aQZzKzde4lBHWsVWT1xcXGYTEYMBoOib9gGg4E7d+7Qr98AFi6cx0svDbH24YSFzaN9+47WpAFQrpwXH344Ntd+3S+//JyPP/48x3N79+7isccey9ZE9dtvS7hw4RxlypQhLS2Nt956l5iYKL79dgoVKlRgzJjPWL16OT/8MJMFC5bg4eHJu++OJCYmmmef7ca+fXsxmYxMnTqLb7/9Cl9fP+Lj4wgIqMaLLw4CoEmTYD744N3iTRxLlixh6tSpdOzYkapVqzJ//nxefTUzK9+4cYMNGzbQp08fmypfvXo1kZGRQOZTjMlksq7A6+vrS48ePazXTp06lVWrVrFw4UKCg4MB0Gq1fPzxx7zzzjsMGDCAPn36kJKSwvz58ylXrhyhoaE2xVPSJBmTcNW4UMXTl/I6b/HNW3CIvSdvsufEzTyvOXctPluPyI6jkew4GokE1PTPfRhoy/qVaFGvkqI4bt6M5OuvJ3LlymUuX/6XV199nTJllDUD7tq1g1atQmjRohXTpn3D0aOHadQo84vQ6dOneOqpFg+Uad48+7EbN67x2WcfWcvk5uLFC/j43EtC//xzgM2bNzB3bhiSJPHjj7OYN+8nRowYRceOnTl2LLNfqEeP51m0aAGQ2b87dux4+vXrScuWrXnllaGsWPEbf/yxEQ8PT0aMCMVsNrN06SJrPT4+Fbl5M5LU1FRcXFwUvS+5sSlxfP3113Tp0gWA33//Pdu6VJ06deLPP/+0qfIVK1Zw8ODBbMemT58OQNOmTbMljtx07twZJycnZs+ezZQpU9DpdDRv3pz333//ke3fyDKs/mA+3P0FfmUq0jOgW4HKZxHfvIWiVM23LNHxaSSlmgCQJHBz1uLj4Wy3OipV8mX06DEYDOl89tkYunbN//Mjy19/befTT8ej1+sJCXmaDRvWWBOHUn5+/owbNwmAX39dmOt18fHx2fYq2r//b4KC6lmfcOrVa8CPP85kxIhR+dbp4eHB448HAtC7d1/Ono1gyZIwUlNTaNu2HS+8MMB6bVZLTXx8XPEljlu3blmTRk5CQkL45ptvbKo8LCxM8bWTJ09m8uTJOZ5r27YtbdvmPxzyUWO2mEkyJRdoKKwg2EuLesqeChZujmDnscynDGRoHOhdJM1Ver0TLVu25scfZ+XaXHS/mzcjuX79Oj/+OAsAgyGdffv28s47H+Dm5kbdukFcuvQvbdpkL3flymV8fCpYP5Dv9+KLL+Van5ubG0ajIdux+4c2y7Js/V2SJOT7RvFlZGSfqa/VZm+GCwysxW+/rWHPnl0sW/Yrq1atYOrUmXdfl+Fu/YUfjKF4cPB/3xwnJ6dsfR1paWnWZieheCQaM4fPlROJQygBElONuLvqqFKhDG2e8CMh5cFh9PbSvn1nDhzYx40b13O9ZuPGdRw5coiNG9fx8cefMWrUe4wa9R4TJkyhQoWKbNu2GYCXXhrCtm1biImJtpa9cuUyH374rnVfopwsXryAqKhbDxyvXNmf2NhY6+/Nmj3FyZPHrfOLwsNP0KxZ5mCfcuW8uH37NgC3b98mLi7v/uU1a1YSGRlJu3YdmDx5KqdPh1vPxcXdwd3dnbJly+Z5DyXUn3/++edKLvzzzz9JTk6mQYMGAOzbt49Dhw7RuHFj0tPTmThxImlpabz44ouFDsqR0tKMKNvaKjtXVz2pqUX3H0JOolNj2Bt5kHbVWuKh9izWurM44nWLuh1bd3R0NG5uypenyNK0dgWOX7iNRq1iRM96NK1tW1OySiXlOMcqPPwEv/22lOvXrxIfH0+TJsFoNBoSEhJYtmwxbm5urFmzkgsXzpOYmMD+/X+zb99eDh8+RLVq1Vi6dBG+vn4EBtYGMuc87N//N0eOHKJChUrUq1efp556innz/o+DB/eze/dfHDy4nw8//JRy5cplq//mzUj27dvLvn17OXToAG3atHugn8XHx4f58/+P559/AUmS8POrjMFgYPXqFRw5coiUlFSGDx+JVqulQoWKbN++hUOH/iExMZFLl/7l5s1ImjRpxvffT+fcubPExETTpEkzVCoVt2/HsHjxfM6ejWD79q307t2XGjVqApmd8q6ubjz1VKsH3sPk5IRs/6aSJOHikvugAsU7AC5YsIBJkybRpk0bfvzxRw4cOMDLL7+c7Zovv/ySXr1sn4T2MClJEwBPxJxizskFTGr/IWXNua9cXJQcPRlN1F38dZ88GU7FigXb8VNMAMy0aNF8nJ1d6N27b7HUnZqawkcfjebzzyfg6fngZ8WtW1ey/ZvabQLgoEGD6NOnD2q1GoDg4GBmzZrFypUr0el0dOrUiY4dO9ryWoRCSjBmjkTydHLHbNuK9oIgONDAgS/zxx8bSU9Pt3l1i4I4fvwYH300NsekURA27QAImZPusmZpt2vXjnbt2tklEMF2CYYkJCTcncpwJ8W2vUwEwREK8qTxqOrYMffBRvb236HDhaW4c/zDDz9kzpw5ogP8IZJoTMRN54papXZ0KIIglCKKnzhcXV1ZuXJljkPPBMdIMCTiriv8CAlBEARbKH7iqFq1ap5DzwCWLl1a6IAE5RKMSbjrReIQBKF4KU4cI0aMYOLEiTkuYZ5l1qxZdglKUCbRkIi7TqysKpQcqesmkbpukqPDEApJcVPVn3/+SXh4OK1bt6ZevXp4eXk9sJFLcrJtu0gJBWeRLSQakykrnjgEQShmihPHqlWrrD/v3r07x2vENpLFJ8mYjIwsnjiEUu/+/Tief/4FoqOjWLNmJV9+OYWQkKeBzL025s6dQ2pqKi+9lLlbaVaZfv0GMGDAvbXT/rt/hyzLXLt2hb59BxASUvqWNsqJ4sRRvnx59uzJe5ezli1bFjogQZmsORyij0MoSWRzBpa4SCyp8ahcbJ99npP798PI2kdj//6/mTTpC2rUqImfX2Vq1qzFCy8M4NatmzRpkrm6dlaZ+5MG8MD+HXq9ngsXIhg27FVWrdqAu7t94i7JFPdx9O/fP99r3nzzzUIFIyiXaMicOVxWjKoSShBL8m0wpWE8vKZI62nUqDFdu/bkk0/+Z13crzACAqphNBq4eTPv5eNLC8VPHEqSwvPPP1+oYATlEgxZTxyiqUpwLNO5vZjO7srzGvPNc9y/R7npzA5MZ3YAEupKNXMtpw1sjbZmwSavDR8+krffPs20ad/wwQcfF+geWQ4e3I+nZzmqVg0o1H0eFXbdc7xt27b5NmcJ9pHVVFVW9HEIJYDKpxqWxGhIz1pjSwInN1RlffIsVxhqtZovvpjEkCED2bx5Az4+tu/PM23a11gsFjIyTHz33fdiHttdihOHkqG2qali2YvikmBMwk3rikZl19wvCDbT1myh6KkgffcC61MGgPaxxji1Gpx3oULy9CzHhAlTeP/9UF57bbjNH/xZfRyOXGDxYWTXxCFGVRWfREOSeNoQShQ5LQFc3FG5eKD2qY6cGl8s9datG8SwYSOYNes73nvvwzyvzWuvcOEexYnD09OT5cuXZztmsVi4c+cOR48eZffu3fzvf/+ze4BCzhIMiWJElVCiOHcItU7+c2qZ+w55tgoPP8Eff2zi1q2b/PLLT8THx3H6dDirVy+nR4/MfteePZ/PtqnR/WWmTv3Kevz06XDi4+OZO3cOADNnTqV//0FUrVrFbvE+ChQnjoEDB+Ln5/fAcX9/fxo0aEBwcDDz589n0iQxK7Q4JBgTqeT6aO+pLghKBAXVZ/r0H/K97v4nifzKfPLJOD75ZJw9wnsk2W1UVe3atdm7d2+hAxLylzlrPImyYkSVUMK4dP3I0SEIdmC3ntVDhw5hsZTOzqPUdZOI1GrQdhpdLPWlmFKxyBaxMq4gCA6hOHF89FHO3xTMZjM3b97k2LFj9OzZ026BCbm7N4dDJA5BEIpfgdaqup9araZChQr07duX9957z26BCblLMGaOhReT/wRBcAS7rlVlK4vFwsKFC1m6dCk3btygXLlydO7cmdDQUFxcXPItbzKZmDt3LmvWrOHatWu4urrStGlT3n77bev2to+iREPW5D/xxCEIQvGz65Ijtpo4cSJhYWG0b9+eIUOGcPHiRcLCwjh9+jTz58/Pc+MoWZYZMWIEu3btol27dgwcOJC4uDh+/fVXXnjhBZYuXUqNGjXsHvPDwLrAoZjHIZQw0478CMDbjYY7OBKhMBQnDiWLHNri/PnzLFq0iA4dOjBz5kzr8cqVKzNhwgQ2bNhA165dcy3/559/smvXLl544QW++OIL6/Hu3bvz3HPPMWHCBObPn2/XmB8WCYYkXDTOaNVaR4ciCA53/7LqrVu3wWAwEB0dxdtvj6ZateocP36UlSt/R6/Xo1KpiI6OpkWLlvTu/UK2+wwe3J9nnunIoEEv51hPWloaU6Z8ycmTx1m+fF0xvLJMK1cuZ+7cn/jsswk0atS42OrNi+LVcbM6v3v27MmxY8esx6Oionj66adZuXKlTRWvX78eWZYZPDj7kgN9+/bF2dmZtWvX5ll+//79APTq1SvbcX9/fxo3bsy+ffuIjIy0KaaSItGYKDZwEkqkDEsG15IiSTAk5X+xQlnLqleq5Mvo0WP45JNxNG/ekokTx3H8+DGmTfua9977kDFjPuPDDz9l1Kj3WLjwl2z3iIg4jcViZsOG3D93nJ2dGTZshN3iVqpXr+fx93+4JiAqfuJYvnw5iYmJvPPOOwQGBlqPe3l5MXz4cKZOnYpOp+O5555TdL/w8HBUKhX169fPdlyv11OrVi1OnjyZZ/msLWydnJweOJd17Pjx4/j6+iqKpyRJEFvGCiVUbHoc6eZ0Nl3aSr9avfIvUEABAY9x5cpl5s79kT59+lO27L0vWlWrBvDee9lHiW7cuI4vvpjM0KEvcezYERo2bJTrvWVZZu7cOdy4cZ3o6CjeffcDqlWrzpEjh1i+fBn+/lWIirpFz559aNCgIQD/938/Eh8fh7OzC9evX+Ozzybg5OTEn39u5dChA5Qt687t29GMGDEKL6/yxMREM3HiONzdPShf3guTyVQ0b1QBKU4chw8fZvbs2dSsmX0JZI1GQ9++falTpw5jx45VnDiio6Px9PREp9M9cK5ChQocPXoUo9GY43mAxx9/HMh88qhVq5b1eFpaGsePHwfg1q1bimIpaRKMSdTweMzRYQgCAAduHmbfzX/yvOZC/CXk+5ZV3x25n92R+5GQ8vxbbl6pCcGVnrQ5pn/+2U/Dhk9w9Ohhhg8f+cD51q3bWH82GNKJj4/nsceqERLyNBs2rM0zccTERNOxYxcqV/Zn584/+fLLz5k7NwydTk9o6HtUrFiRuLg4Ro4cyuLFy0lMTOC335awadN21Go1K1b8htmcwdWrl5k372cWLlyKSqVi3brV/PDDDD799AumT/+Gpk2b07//QFJSkujZU9nnanFRnDiSkpIeSBr3CwoKIjo6WnHFaWlpuSYFvV4PQHp6eq7XdOvWjdmzZzNjxgxcXFxo3rw5cXFxzJw5k/j4eGsdtvLycrO5TKQ282309i76pwBZlkk0JlHRo3y2+oqj7tyIuktX3SqVCo3mXiu3Si2R3/qmj7n7E5MaS5IpBQAJCTetC94uXnmWVamlbHXd/3P2mCRu3ozk228nkZ6ejqurK5999gU9ez6HWq3KtRzA1q07aNOmLRqNiq5du/H++2/z/vv/w9XVjRkzvmPXrh0AzJnzC2q1Ck/PcgQEVAWgcePGfPLJBxgMqfj6VmT+/F9QqVSo1SquXbuKRqPCw8OdOnXq8PrrL9O587O0b98Jd/eybNmyEZPJyHffZa6VlZKSislkRKNRceTIIQYPHoJGo8Ld3Z1q1arl+zoKQ6VS2fQ3ZdPM8eTkZNzccv5gTUqyrc3S2dmZ2NjYHM9l7diVUzNUFnd3d+bNm8cHH3zAp59+aj3epEkTXnvtNWbPnp1rrHmJjU3GYpHzv/A+xvR05ISbRF25ZrftMHOTYkolw5KB1qwnJibzPff2LmP9ubiJuktf3Zn7U9xbJaKJTyOa+OT+DT3LkoiV7Ln7lAHQ0LueouaqrLryWtrcYpGpVMn3gSaoOnWCuHjxIoGBdbIdv3DhPAEBj6HRaNi69Q98fCpy8uRJZFlGo9GwefNmunfvxYgRoxgxYpS17ps3IwHZGofJZLHGOGHCOJo0aUb//gMBWLZsifW6adNmc+bMKf74YyP9+/dm9uxfMJstVK5cJVvMqampZGRYkGUwmzPr0WhUd3+3FNnS7haLJdvflEol5fklWnH6Cg4O5oMPPiAxMfGBc7GxsYwZM4bg4GDFgfr4+BAXF2ftq7hfVFRUrs1Y9wsMDGT16tVs2bKFRYsWWf8/657VqlVTHE9hWJJvIxtSi3w7TLhv1riYwyGUMEnGJNx1ZfAv40tLv2YkGos++Q0d+gYrVvyW7YvtkSOHmDLlSzQaDTduXKdyZX/ef/9DRo16j7fffp8BAwbn2Ul+584dbty4DsCxY4cJDKyNq6sbCQkJuLu7A9mbyW/fvs3ixQuoXbsub789mvr1n+DKlUs0adKMs2dPk5qa+RR27lwEM2dOBeDJJxtz5MghIPNL+aVL/9r3jSkkxU8co0aNok+fPrRq1Yp69erh4+ODyWQiJiaG8PBwXF1d+f333xVXHBQUxJ49ezhx4gSNG98bYmYwGIiIiMh2LD9Vq1alatWq1t93796Nm5sbjRrl/y2oMJLmDgXzvU4r63aYai1lXv25SOq0zuEQo6qEEmZY/cHWeRz9Au23PNH9S6RPm/YNb7/9vvVcUFB9/ve/j5kx41ucnZ0xGo1YLBa++moqMTHRTJ48HldXVwyGdPT6zBaOf/+9yOnT4cycOZW33nrXeq+0tDTmzp1D+fLerF+/htjY20RG3rCuujt8+JvMnj2TiIjTODllbhg1d+4c+vTpz7FjR4iOjkKSVJQtW5ZmzVqg1Wp5++3RjB8/Fl/fyiQnJzFiRCgAo0a9z8SJ47h48Ryenp5UruzP778vxde3MhUrVrTbe1dQkizLittlLl68yPjx4zlw4ABZxSRJolmzZnzyySc2zdY+e/Ys3bt3p3379tnmcYSFhTFhwgSmTJlC9+7dgcyO9KSkJHx9ffPdwSur/MiRI3nrrbcUx5PFlqYqS2o8hn1LybiYOTQYtRbNY0+ib9avyJqsDtw8zMIzy/is2f/wcSkPOL7pQtRduuo+eTKcihWr5n9xDgozAdCRu/A96nXfunUl279pfk1VNvVxVK9enfnz5xMXF8e1a9eQZZmqVavi4WH7h2RgYCADBgxg0aJFjBw5kpCQEOvM8aZNm2ab/Dd16lRWrVrFwoULszWHDR06FH9/f6pXr44kSezdu5dt27bRpk0bhg8v+pmpKhcPJN19icoPm6kAACAASURBVMxsQtI6F2k/h1jgUBAERyvQsuqenp54enoWuvIxY8bg5+fHsmXL2LlzJ56engwcOJDQ0NA8lxvJ0rBhQzZt2mRdgLFatWqMHTuWfv36oVarCx2fElnbYUrmDGSTAUtKXJHWl2BMxEnthF6dd/+PIDyMxFIjjwbFTVXnz59n+vTpAIwcOdI6dyIqKop33nmHt956i+bNmxddpMWkIKOqUtdNQm02YIy+jL7lS+jqPF1E0cH/hS8iMvkmY5vd2/vD0U0Xou7SVXdhmqoK41FvLnJk3bY2VSkeVfX7779z8OBB6tevn61zxsPDgyZNmhAaGmr31XNLEknvgqp8AKbwbdjQbWSzBEMiZcWsccGhJGS5dG7a9igqyOeV4sTx999/M3v2bIYNG5atT0Ov1/POO+8wbdo0vv/+e5sDeFRIkoQuqD2W+EjMN04VWT2JhkTRvyE4lE7nRHz8bTIyTEX6JUkoerIsk5KSiEZjW9O34j6O+Ph4nnwy96n/LVq04MMPP7Sp8keNpnpTpAPLMJ7cgqZykN3vL8syCcYkMYdDcChPT2+SkxO4cycKi8VcbPWqVCqHbU/9KNet0ejw9PS2rYzSCy0WS55rRxkMBszm4vsjehhJai3aOk9jPLwaS/xNVB6V7Hr/dHM6JouJsmLnP8GBJEmiTBkPypQp2lUS/svRfTulse7cKG6qql+/PhMnTswxOZhMJqZMmfLASrelkbZ2W1BpMIZvs/u9xaxxQRAeBoqfOEaOHMmAAQPYvn07zZs3zzZzfO/evaSlpfHrr78WZawlgsrFHU2NYEzn9qBv0gtJ72q3e2ftYSD6OARBcCTFiSMoKIiffvqJsWPHsmZN9jWZqlatysyZM6lbt67dAyyJdEEdyDi3F9PZ3ejqd7LbfcWWsYIgPAxsmgAYHBzM5s2bOX36NFeuXAEgICCA2rVrI+W3rnIpoi5fFXXFmhhPbUMb1AFJwWRGJbKaqsTuf4IgOJLNn2iSJFG3bl26dOlCly5dqFOnjjVp5LVHeGmjrdcBOek2GVeO2u2eicYkdGodTmq93e4pCIJgK7t8FY6Li2PRokVcv37dHrd7JGiqNkJy88IUvsVu98zaMlY83QmC4EgFWqsKMoffbtu2jbVr17J3714yMjJK7QeaS9ePHhgyJ6lU6Oo+g+HAMsy3r6AuX/glGhKNSZQVI6oEQXAwmxKHLMv8/fffrF27lm3btpGamoosy7i7u9OxY0c2bdpUVHGWSNparTEcXoUxfCvObV4r9P0SDIlULuNrh8gEQRAKTlHiOHXqFGvXrmXDhg3ExsYiyzJ6vR69Xs/06dNp0aIFGo2GixcvFnW8JYqkd0VbsyWmiF1Ygvuici7c00KCMZG6+lp2ik4QBKFgck0c169fZ926daxbt45Lly4hyzIqlYrg4GC6detGhw4d6NSpEyEhIdYyixcvLpagSxJdUHtMp7djOrMDfaPuBb5PekY6BrNRTP4TBMHhck0czzzzDJIkIcsyderUoWvXrjz77LP4+PgUZ3wlnsqjEmr/ephObUfX4FkkdcG6lRLu7s8sVsYVBMHRch1V1apVK9RqNYGBgbz++usMGDBAJI0C0gV1QE5LIOPfgwW+R6LY+U8QhIdEronj559/ZteuXfTp04dffvmFFi1aMGbMGPbt21ec8T0S1JWDUHlUwhi+tcDLUGc9cYjEIQiCo+U5j6NcuXIMHDiQZcuWsWLFCnx9fRk3bhytW7fmq6++emDBw+3btxdpsCWVJElog9pjibmEJepCge5hfeIQTVWCIDiY4gmAVapUYeTIkWzevJlZs2aRkZGBWq2mV69ezJ8/n+joaMaOHVuUsZZo2sdbgM4FYw4TAlPXTSJ13aQ8y8cbE9GqNDhrnIsqREEQBEUKNHO8fv36fPzxx+zatYt3332X06dP06VLF+Li4uwd3yND0urR1goh49JhLMmxNpdPNGRO/iutkywFQXh4FGrJEZVKRcuWLZkyZQq7d++mTBnRjJIXXd12gIzp1J82l00wJuEuNnASBOEhYJ9lWwFnZ2e++OILe93ukaQqUx5NwJMYI/5CzjDYVDbBkCiWGxEE4aFgt8QB0KFDB3ve7pGkDWoPhhRM5/62qVyiMVGMqBIE4aFQ4EUO7cFisbBw4UKWLl3KjRs3KFeuHJ07dyY0NBQXF5d8y8uyzPr161m8eDGXLl3CaDTi6+tL586defnll3FzcyuGV2EbdcWaqMpXxXRqK9rabRT1WRjNJtIy0sWIKkEQHgp2feKw1cSJE5k0aRI1atTg008/pVOnToSFhTF8+HAsFku+5adNm8b777+PXq9n5MiR/O9//6NmzZrMnDmTIUOGFHjORFGSJAldUAcscZGYb5xSVCbRKDZwEgTh4eGwJ47z58+zaNEiOnTowMyZM63HK1euzIQJE9iwYUOeG0NlZGSwYMEC6taty7x581Dd3WWvf//+qNVq1q1bR0REBLVr1y7y12IrTfWmSAeWYQzfiqZyUL7Xx9+dw+Eh+jgEQXgIOOyJY/369ciyzODBg7Md79u3L87OzqxduzbP8hkZGaSnp1O+fHlr0siStTSKs/PDOedBUmvR1m6L+epxLPG38r0+MWudKjGqShCEh4BdE8fvv/+u+Nrw8HBUKhX169fPdlyv11OrVi1OnjyZZ3knJyeaNGnC7t27+emnn7hy5QrXr19n5cqVLFmyhG7duhEQEFCQl1EstHXagkqD8dTWfK9NsM4aF08cgiA4nl0Tx/Tp0xVfGx0djaenJzqd7oFzFSpUIC4uDqPRmOc9vvnmG4KDg/n222/p0KED7dq1Y8yYMQwePJgpU6bYHH9xUrl4oKkejOnsHmSLOc9rEwyJqCU1rtr8BwwIgiAUNcV9HOnp6UydOpVt27YRHR39wDpVtkpLS8sxaUDmU0dWnbldA6DVavH396dChQq0atUKSZL4448/mD17Nnq9njfeeMPmuLy8Cj4Sy9vbtqYkQ6se3Di/FyktHoshFU/nDDRung9e928ans7u+Pjk/sRha932JOoWdYu6H926c6I4cUycOJGVK1fSoEEDnnjiiQc+0GVZZvPmzYordnZ2JjY256U3DIbMyXFOTk65lk9LS6N///7UqVOH7777znr82Wef5Z133mHGjBl07NiRatWqKY4JIDY2GYvF9tFY/91zXBGNN+qKNTHfOg/I3NyyGKdWgx+4LDrhDm4at1zvX6C67UTULeoWdT96datUUp5fohUnju3bt/PLL7/QtGnTXK/Zs2eP4sB8fHy4cOECRqPxgSQUFRWVazNWlj/++IPLly/z7rvvPnCuU6dObNy4kcOHD9ucOIpT0tyhYDZZfzed2YHpzA5Qaynz6s/W44nGJLydvRwRoiAIwgMU93FIkpRn0gDbllUPCgrCYrFw4sSJbMcNBgMREREEBeU9TDUqKgogxyazrGOFbU4raq79v0ZTPfjeAbUWTY1muPb/Ott1CQYxa1wQhIeH4sTRrFkzIiIi8rzm//7v/xRX3KVLFyRJYsGCBdmO//bbb6SlpWWbwxEdHc3FixdJS0uzHqtevToAq1evfuDeq1atAqBevXqK43EElYsHku6+Dm+zCTROqFw8rIdMlgxSMlLFOlWCIDw0FDdVDRkyhK+//po2bdrQsGFDPD09H5g/sWjRIkaMGKHofoGBgQwYMIBFixYxcuRIQkJCuHjxImFhYTRt2jRb4pg6dSqrVq1i4cKFBAdnfkNv27Yt9evX56+//mLAgAF06NABWZbZunUrhw4dolOnTtStW1fpy3MYOS0BXNyR1DrkpJgHZpMnGrJ2/nu4OscEQSi9FCeOnj17AvD337YtzpeXMWPG4Ofnx7Jly9i5cyeenp4MHDiQ0NDQB5LSf6nVaubNm8dPP/3Eli1b+Prrr5EkiYCAAN5//31eeeUVu8VZlJw7hFo3cVJVbYgpfCumS4fQPtYYgISs5UbEOlWCIDwkFCcOZ2dnXn311VzPy7LMvHnzbKpcrVYzZMgQhgwZkud1kydPZvLkyQ8cd3Nz4913382xg7wk0ge/gDnqAul/zUXtVQVVWZ97W8bq3R0cnSAIQibFicPV1ZWRI0fmeU1+y4QIeZPUGpyfGUHKis9I2/YDLt0/JsEomqoEQXi4KO4c37ZtW77XbN2a//IZQt5UZbxxavMaltuXMexbSqIhEZWkwk3r6ujQBEEQABueOLIm48myTHh4OFevXgWgatWq+Q6dFWyjDWiEuV5HTCf/4I62KWV1ZVBJDl0BXxAEwcqmZdUPHDjAp59+yrVr17Idr1KlCuPHj893noegnD64D+aoC8TfvkBZLz9HhyMIgmClOHGcPHmSoUOH4uLiQpcuXaxLl0dHR7N3715ee+01fv31V/H0YSeSKrO/I2nPRDziY5AzjEia3GfSC4IgFBfFiWPmzJk8//zzfPjhhw8sBWI0Gpk8eTIzZszgp59+snuQpZXKzYskZxf878Rj2PcrTq1ednRIgiAIyjvHT506lWPSANDpdHzwwQeEh4fbNbjSzmwxk2w24Fm+OqYzOzFd2O/okARBEJQnDlmW81x0MGspdMF+snb+83wsGHXFmqTvno8l/qaDoxIEobRTnDj8/PxYs2ZNrufXrFmDr6+vXYISMmUlDg8nd5zavYGk1pK27XvkDIODIxMEoTRT3Mfx8ssvM3r0aDZv3kybNm3w9vYGMjvHd+7cye7du/n666/zuYuQE5euH+V4PP6+LWNVrp44tR1G2qapGPYuxikk79n2giAIRUVx4nj22We5ceMG06dPZ+fOndbjsiyjVqsZNWoUXbp0KYoYS63ErHWq7s4a1/jXQ/fEcxiPrkNdKRBtzRaODE8QhFLKpnkcw4YN49lnn2XLli1cuXIFWZZ57LHHaN++PX5+Yq6BvSUYkpCQKKO9txOX7skemG+dI33PAlTeAag9xfsuCELxsilxQGZfR24rz5pMJrRabaGDEjIlGBIpo3NDrVJbj0kqNU7t3iB1xVjSt32PS4/PALGOlSAIxceu61i0bdvWnrcr9RKNibjnsJy6ysUDp6eHY4m7SfrehQ6ITBCE0izXJ45z586Rnp5O/fr1AZg1a1a+N0tNTbVfZAIJxiTK5rJlrMavDrpG3TAeWUPS8Ybg26SYoxMEobTKNXEMGjSIlJQUDh06hJOTk6LEIUmSXYMr7RIMifi75T7EWdeoO+ao89ze/DPOPSqiLudfjNEJglBa5Zo43n//fZKSkqyr4np6erJ8+fJcbyTLMn369LF/hKWURbaQZEzGPZcnDgBJpcKp7eukr/qM9K3f49LzMySdczFGKQhCaZRr4vhvEnjqqafyHTn11FNP2ScqgSRjMjIyZXW5Jw4AlYs7Pj3f4ebiz0nfswCntq+LJz9BEIqU4lFVOW3dmmX//v2kpqby7bff2iUo4d5e40p2/nOuGoTuyZ4YD63EVKkWutptijg6QRBKM8WjqvIaMXXmzBnGjBnD+PHj7RKUkNm/AeTZVHU/3RPPoa4chOHvRZhvXynK0ARBKOXsMhz3lVdeYcuWLYq2lxWUSTTc3Ws8n6aqLJKkwqntMCS9G2nbfkA2phVleIIglGJ2SRwWi4Vbt26RkZFhj9sJ3GuqKqNzy+fKe1TOZXFq9wZyUgzpu+Yhy3JRhScIQimWZx9HrVq1snW01q5dO8+bderUyT5RCSQYEnHTuqJR2Ta5X1MpEF2TXhgPLsdUKRBd3XZFFKEgCKVVnp9KPXr0QJIkZFlm8+bNdO7cOcfrnJ2dqV69Or169bKpcovFwsKFC1m6dCk3btygXLlydO7cmdDQUFxcXPIse+DAAV566aU8r/n111958sknbYrpYZFgTFLcv/FfugZdMN88h2HfEtQ+1VF7B9g3OEEQSrU8E8f9I6n27NnDpEmT7Fr5xIkTCQsLo3379gwZMoSLFy8SFhbG6dOnmT9/PipV7i1p1atXZ8qUKQ8cNxqNjB07Fk9PT+us95Io0ZBE2RyWG1Eis79jKKkrPiNt2/e49h6HpMs7EQuCICiluB3E3h3f58+fZ9GiRXTo0IGZM2daj1euXJkJEyawYcMGunbtmmv58uXL07179weOr1+/HovFQvfu3Uv0gosJxkQquVYocHmVUxmc271B6rrJpO+ci1P7kWJ+hyAIdqG4c1ytVhMREUFERES2NalSUlLYt2+fzRWvX78eWZYZPHhwtuN9+/bF2dmZtWvX2nxPgN9//x14cAJjSWKRLSQWoqkqi7ri4+ibPk/G5cOYTokRb4Ig2IfixLF+/Xp69OjBwIEDuXDhgvV4eno6w4cPZ/DgwSQnJyuuODw8HJVK9UBzkl6vp1atWpw8eVLxvbJcu3aNAwcO8OSTT1KtWjWbyz8sUkypWGSLdQOnwtDW74S6SkMM+5dijv73gfOp6yaRus6+TZCCIDzaFDdVrV+/np49ezJu3Dh0Op31uJeXF3v37uWTTz5h+vTpfPzxx4ruFx0djaenZ7Z7ZalQoQJHjx7FaDTmeD43K1asKPSaWV5eyoe//pe3t332xUiJSwDAv7yP4nvmdZ35+be5MXc0xh2z8Xv1G9TO915jpFaTb/nC1F3URN2iblF38VOcOP79919mzpyZ4we5m5sbX3zxBT179lScONLS0nJNCnq9Hsh8mlGaOMxmM6tWrcLNza1Qw4JjY5OxWGyf/+DtXYaYmKQC13u/y7E3AZAMWkX3VFK3ru1wUtdO5MaKaTh1CLX2d5hMmXNvChq7PV+3qFvULep+OOpWqaQ8v0QrbqoymUx5DpEtW7YsBoNBcWDOzs4YjcYcz2XdJ2tlXiX27NnDrVu3ePbZZ3F2LtkrxCbYOGtcCbVPdfTBfcm4chTTyT/sdl9BEEofxYnDzc2NEydO5Hr+5MmTuLq6Kq7Yx8eHuLi4HJNHVFRUrs1Yucla8r0kd4pnSbw7a7ygw3Fzow3qgCbgSQwHfsccdSH/AoIgCDlQnDi6devG8OHDWbJkCdeuXcNoNJKcnMylS5eYN28eb7zxBj179lRccVBQEBaL5YFkZDAYiIiIICgoSPG9YmNj2bFjB4GBgdSrV09xuYdVgiEJF40zWrV9hxNLkoRTyBAkt3KZ61mlKx/MIAiCkEVxH8fQoUM5duwY48aNe2A+gCzLhISE8NprrymuuEuXLsyZM4cFCxbQuHFj6/HffvuNtLS0bHM4oqOjSUpKwtfXN8dmqNWrV2MymR6Jpw3InMNR2KG4uZH0rjg/M4LUNV+StvNnLBkm5PibWFLjUbl4FEmdgiA8WhQnDq1Wy5w5c1i7di2bNm3i6tWryLJMQEAAnTt3pmvXrjZNMAsMDGTAgAEsWrSIkSNHEhISYp053rRp02yJY+rUqaxatYqFCxcSHBz8wL1WrFiBXq+nW7duiut/mCUaEu3av/Ffau/H0Dfrh+HvRaDVg8mA8fAanFoNzr+wIAilnk0r6EmSRPfu3XOcsV0QY8aMwc/Pj2XLlrFz5048PT0ZOHAgoaGheS43cr8jR45w8eJFnnvuOdzd3e0Sl6MlGJOo4VG+SOswHFiW+YMpcyCC6cwOTGd2gFpLmVd/LtK6BUEo2WxbepXM0VWnTp0iJiaG9u3bk5qailarLdDyHmq1miFDhjBkyJA8r5s8eXKuOxA2atSIs2fP2lz3w0qW5SJ/4gBw7f81hr2LyLh0yHpM8qqCc7s3irReQRBKPpv241i5ciUhISH079+fDz74AIDjx4/TunVr61IfQuGkZKSSIZuLrI8ji8rFA8kp+6gtOfYqqau+wHDwdyyp8UVavyAIJZfixPHnn39am5b69etnfcJo3Lgxn376KdOnT2fLli1FFmhpkbXzn72H4uZETksAF3dU5auirfM0Kt/aaPzrYTy+kZQl75O+ZyGWxOgij0MQhJJFcVPV//3f//Huu+8ybNgwALZu3Qpkdpp36dKFSpUq8dVXX9GhQ4eiibSUyNr5r6ifOACcO4Ra16lyanlvbxNLwi2MxzdhitiF6cwONNWC0TXsgtqrSpHHJAjCw8+mJUfCwsJyPf/EE09w69YtuwRVmiUY7iaOIu7jyIvKvSJOrV9B92QPjCe3YDqzg4yL+1H710fX8Fk0lQIdFpsgCI6nOHHkt3+1wWAgPT290AGVdllNVe52WBm3sFSunjg1ewH9E89hPPUnpvCtpK2bhLrC4+gaPou6SsndKEsQhIJT3MdRs2ZNfvnll1zPz5gxg1q1atklqNIswZiIk9oJnVr5citFTdK7om/UDdcXv0HfYiCWlDuk/TGN1OVjSTr5F7Ilw9EhCoJQjBQ/cQwfPpyhQ4eyc+dOWrRogcFgICwsjFu3brFt2zauX7+eZ2IRlEkwJD4UTxs5kTR6dHWfQVu7DRkXD2I8toGYtTOQypRHV78T2sBWSBq9o8MUBKGIKU4cLVu2ZMqUKYwfP54jR44AmXuGy7KMu7s733zzTY6zugXbJBiTHNq/oYSk0qB9/Ck0NZrhmnCemL9+x7B3EcbDa9AGtUdXtx2SXvmCl4IglCw2TQDs2rUr7dq1Y8+ePVy5cgWAxx57jKeeeirPJdcF5RINiQS4F9/oJZeuHxW4rCSpcH28MakegWTcPIvx2AaMh1ZiPL4Rbe226Op1QOXqacdoBUF4GNg8c9zFxUUMuS0isiyXiCeOnGgqBaKpFIg59irG4xsxndyMKXwr2pot0DXojMq9oqNDFATBTmxOHELRSctIx2QxFcscjqKi9qqC89PDsTTuhfHEZkxnd2E6uwvNY00y54KUD3B0iIIgFFKuieOjjz4iPj6eH374AUmSaNeunaIbqlQqKlasyHPPPccLL7xgt0BLg6wNnNyLYdZ4UVOV9cGp5UvoGnXDFL4V46ntZPx7EHXloMyhvJVq2bSasiAID49cE8exY8dISkrCZDKh0+mIioriiSeeUHTT+Ph4xo0bR1JSkk17dJR2WVvGli3BTxz/pXLxQN+0D7qGz2I8vQPTyT9IW/8VKp9qmZMJqz6BJNm0ZJogCA6Wa+JYs2YNZrPZun2rh4dHnjPH/+vYsWN88MEHInHYoDiXGyluks4FfcNn0QW1x3RuD8bjm0jfMhOVRyV0DbqgqdEcSS1aTgWhJMj1v9T/7vc9bdo0m27csGFDkpPF1qS2uLfcSMlvqsqNpNGhq/M02lohZPz7D8bjG0j/ay7SoVWZc0FqhSBpxVwQQXiYKf6Kd//2rrIsc/v2bQDKly//QFv19evXWbJkCV5eXnYKs3RINCahU+tw0jg5OpQiJ6nUaGs0Q1M9GPO1kxiPb8Cw71eMR9bemwvi5OboMAVByIFNbQOxsbF88803bN26lZSUFABcXV3p2LEj7777rjVR7Nixg8uXL/Pqq6/aP+JHWIIh8ZF+2siJJEloqtRHU6U+5lvnMRzbgPHwqrtzQdqgq9cRlVs5R4cpCMJ9FCeO6Oho+vbty61bt/D29qZq1arIskxMTAwrVqxg3759LFu2DG9vbwYNGsSgQYOKMu5HUoIx8ZHs31BKXfFxXDq9jfnO9cy5IOFbMZ3ahvbxp9A16ILKo5KjQxQEARsSx/Tp06lUqRKzZs0iKCgo27mTJ08yadIkZsyYwfjx4+0eZGmRaEjCv4yfo8NwOHW5yji3HYalcc/MuSARuzCd3YMmoFHmUF6fao4OURBKNcWJY8+ePaxcuTLHfot69eoxffp0evfubdfgSpsEYyJ19WKF4SyqMt44tRiErlH3u3NB/iTj8mHUfnXQNXgWubxYG00QHEFx4jCbzXl2dnt7e2OxWOwSVGmUnpGOwWwskcuNFDWVc1n0TXqja9AF05mdGE/+QdrGr4k8Wh2pbic0AU8iqcRcEEEoLooTh6urK6dPn6ZOnTo5ng8PD8fVVayIWlAJxqwNnETiyI2kc0bXoDPaoGcwnduLOXwzGdu+R3KvmHn88aeQ1FpHhykIjzzFiaNbt24MGzaM1157jZCQELy9vYHMTvO//vqLuXPn8uKLLxZZoI+6xLtzOMqWslFVBSGptehqt6F8yy7c/GcnxqMbMOyah/HwanT1OmbOBdE5OzpMQXhkKU4cw4YN4/jx40yePJmvvvoq2zlZlmndujXDhg2zqXKLxcLChQtZunQpN27coFy5cnTu3JnQ0FDFy7RnZGTw66+/smrVKi5duoRaraZKlSq88MIL9OvXz6Z4HEk8cdhOUqnRVmuK5rEmmG+cwnhsA4b9SzEcXYeubju0dZ9B5Zzz+5m6bhJQuGXlBaG0Upw4tFotc+bMYe3atWzatImrV68iyzIBAQF07tyZrl272rxo3cSJEwkLC6N9+/YMGTKEixcvEhYWxunTp5k/fz6qfNqtjUYjb7zxBgcOHKBr167069ePjIwMrly5QmRkpE2xONq9WeMicdhKkiQ0lYPQVA7CHH0R47GNGI+sxXh8M9raIejqd0LlJiajCoK9KE4c//zzDwDNmjWje/fuha74/PnzLFq0iA4dOjBz5kzr8cqVKzNhwgQ2bNhA165d87zHDz/8wL59+/jll19o1qxZoWNypARjIlqVBudSMGu8KKl9quPc4S3McZGZc0FObcd0ajuaGs3QNeiCulzmcGfZnIElLhJLajwqFw8HRy0IJYvioSiDBg1i5MiR1m1jC2v9+vXIsszgwYOzHe/bty/Ozs6sXbs2z/KpqaksXLiQdu3a0axZM2RZLtFrYyUakiirKyuWGrcTtacvzm1ew7X/FLR125Fx6R9Sl39M2h/TMUddwJJ8G0xpGA+vcXSoglDiKH7icHJyYuPGjXZbfyo8PByVSkX9+vWzHdfr9dSqVYuTJ0/mWf7QoUOkpKRQt25dJkyYwIoVK0hNTcXT05O+ffsSGhqKRlNyVltNMJTuWeNFReXmhdNTL6Jr1BVT+DaMR9aQceWo9bzpzA5MZ3aAWkuZL2nBzwAAIABJREFUV392YKSCUHIo/mT19/enbNm8P9j++usvQkJCFN0vOjoaT0/PB1bhBahQoQJHjx7FaDTmeB7g0qVLACxYsACtVsvo0aPx8PBg3bp1zJkzh6ioqAc68ZXw8ir4wnre3gUfEZViTsHf3bfA9yhM3YVVMuouA/4vYQx+huhV32G89W+2syonN8zbZ6KrEICuwmPoK1RF41kxz71CSsbrFnWLuu1PceIYNGgQc+bMYeTIkble8/HHH7Nnzx5F90tLS8s1Kej1mctqp6en53pN1iKLCQkJrFu3jurVqwPQpUsXBg0axOrVqxk6dCg1atRQFE+W2NhkLBbZpjKQ+Q8bE5Nkc7ksd9LiqeFevUD3KGzdhVHy6i6D7FkVbv0LSICMyqc6qrI+pMdeJfXiUZDvTmTV6FF5+aP2qoKqnD/q8lVQlauMpNGXwNct6hZ1K6dSSXl+ibZpkcMtW7awefNmGjdujJeX1wPt8ampqYoDc3Z2JjY2NsdzBoMByGwey03WuQYNGliTRpYePXpw8OBBDh48aHPicASj2UhaRnqpWxnXUeS0BHBxR+XigdqnOnJqPM5Pv555LsOY2WkeexVz7FUsd65hOr8PTNvvlpZQuVfA4lsNk5svai9/VF5VkFw8RP+UUGooThyzZs2y/nzhwoUcr7HlPxwfHx8uXLiQY3NUVFRUrs1YWSpWrAhgnYh4v6xjiYmJiuNxpKwtY0UfR/Fw7hBqncfh1PKlbOckjQ61dwBq7wCy5qDLsoycfDszkdzOTCaGyAtkJPx9r5xTGVReVe49oXhVQeVREUlVcvrZBEEpxX/Vnp6eLF++PNfzsizTp08fxRUHBQWxZ88eTpw4kW2TKIPBQERERLZjOalXrx4At27deuBcVFQUQInZSMq6ZayYw/FQkiQJqYw3qjLeEPAkkNl8EH39FuY717HEXr37hHIN06ltmMwZmQVVGlTl/FCVq3K3mcsftZc/kl4szSOUbIoTxzPPPIOfX95Lfvft21dxxV26dGHOnDksWLAgW5L47bffSEtLyzaHIzo6mqSkJHx9fXF2zlxKwt/fn0aNGnH06FFOnTpF3bp1gczFGH/77Tc0Gg0tWrRQHI8jJd6dNV5WL5qqios9ZoxLelc0lQKhUqD1mGwxY4m/hSX2CubYa5kJ5eoxMs7tvlfOzeveU4lXFdReVZDKPLiTpiA8rBQnDiX7bLzzzjuKKw4MDGTAgAEsWrSIkSNHEhISYp053rRp02yJY+rUqaxatYqFCxcSHHxvKe1PP/2UAQMG8MorrzBo0CA8PDzYuHEjJ06c4M0338TX11dxPI5knTUumqpKPEmlRl3OD3U5P7SPZx6TZRk5LQHL7auY79xr7sq4egzkuwMxtM53+0v8rclE5emHpMm9uVYQHMXmBtidO3eyfft2bty4AYCfnx/t2rVTPAz3fmPGjMHPz49ly5axc+dOPD09GThwIKGhofkuNwJQp04dlixZwrRp01iwYAEGg4Hq1aszadIkevXqZXM8jpJgSEQjqXHVKFufSyhZJElCcvFAVcUDTZV785bkDAOWOzcy+07udsabzu0F0593C6pQeVTMfDK5r7kLxJOp4FiSLMuKxp6mpqYSGhrK3r17+W8RSZJo2bIlM2bMsDYllVSOGI678PQyzsVdZEKLMQUqX9qGCj7KdcuyBTkxJlsyscReQ065Y71G7eoBnpWzNXep3Csq2pOksIs7Porvuaj7QXYbjvvNN99w5P/bO+/wKKr3b9+zJdn0CiEJRUATWjBACCUoRUiQ3hRUQAR/AhddBAuCKAqCgn4pIvhSJCBRei+CGBEFpJdQQ28hBdI2yWZ35/1jkyXLJiEbktDOfV25Zue058zuZD5z6nPoEH379qVFixZ4e3sDkJCQwK5du1i1ahXffvst48ePf/haP2OIVeOCXCRJgeTmg8LNB6o1NIfLmWkYkq5iTLiCWnsT7fVYdMe3gdFgSqC0Q+HpnyMmlVB4VUHpWdFqe3mxR5egJCiycGzevJmffvqJBg0aWIS/8MILNGnShPDwcIYOHSqEoxgkZd4hXa8lOSsVNzFALsgHSeOMyq8m+NU0v4HKBj3GuzcwJl41t1CyLx6A09H38rmWR+lZCYV3ZZSelTGm3Dbv0aV56e1CLAoEBVNk4VCpVFaikZeQkBDU6mfT+1pyVgqz//iJPgG9ivXgT8q8g142sOXi7/Sq8eSMzQgeLZJShTJnIF2NaQahLMvI6UkWYmJIvIr+0kGLvOY9upBQ13jJtIDRwc00FuNoOkqObsKjoiBfiiwcFStW5ObNm/j6+uYbf+PGDapWrWoRtmnTJtq3b/9wNXwC2HJxB6fjz7NFnf+D32A0kJqdRqoujRRdGqm6VFJ1aayP3YKRe+Mpu2/sZfeNvagUKv7XYnJZXoLgKUGSJCRnLxTOXqiqBJvDDXdvkbUnEsP1kzkJFeDgimTvhP7yEeSMVCCfsT17p3tC4uBGond5dDiaRCVn9b3k6A5qjZhO/AxRZOH44IMPeP/99xk7diz16tWziDty5Ahz5szh008/tQifMmXKUy0cI/78BL1Rbz7PffBLSDzvXpXU7HRSdamkZ+e/FYtKUqGSJLKNemRk1Ao1weXq0PX5DmV1CYJnBKV7BRSu5XOEQwJZRl2lnrm7SjYakDNTkbV3kbV3MWqTcz4nm8+NyWdJvvgfGPTWBlR2Oa2VnJaKVeslJ1zjXOjGkaWF8PhYshRZOMaOHUtycjJvvvkmarXavCo7MTGR7Oxs3NzcGDx4sEWeO3fulGxtHzO+aPIRq89v5PDt4xhk0yClAgUeGjcMspEKjuV4wb0aLnbOuNo542LnYjqqXXCxc0ajsmf56dXsubEPlUKF3qhHo7QX4xyCUiG/PbpykRRKU8vhAQPm3t7OxF+7dU9YMpLvE5q7GBKvImtPQHaGdQGSEsnR1dyCMQtNzvHeuZvYrqUEKC3BtGmTw/tbGmDqwiosz9OMm70rGqUGo2xErVChNxoI8wu1aZwiVZdKM//GNPNrxN839pGS9WTsryV48nAIH/7QZUiShKRxRqlxBs/Cd5KQs7NMCx9zBMXUeslznpaI/nYscmb+U00ljcs9UXFwI9G7XE432X0io36w18xndTZZaV13kYXD3d2dyMhImwpv1qyZzRV60sh98Hes3YoNJ/+w+cH/Xt17M1t6BXYt6eoJBI8MSW2PpC6PwrV8oelkox45I79usrs5QpOM8c4NkmP3gTGfbjK15p6QOOTXenHHmBoP2RlkHViDJqwPSBJIiqd+XCavp8uSnEVX5AWA//77L02aNLGp8OLkedQ8Kn8cD4OwLWw/C7a9vZ24fe1W/q0XrWWXGfqsohcsKcxCcv9RyvmsUCoxynnSIoFCgZRzBAUoJNMxTxlSIWXfi78/zjJe42BPZpbeFGa2d+8o3XeOpEB3aN09vzJ5KaKnyxJbAHi/AGzZsoVXX33VpjwCgUBQXCRJgULjAhoX8KxUaFo5O9M03pJ0Hd2xrRjjzpkiFEokdz9UleogqexND1dZNh9li/N7nzX2KjIzsky7ZpjDrdOZ8yOD0Wh5NBiQkfNPb2X/XphWkjEaDPnUL59jfjPjAJR2qKrWx75xrxL5LYo9+vTVV189UDgEAoHgUSCpNUhuFVC4VcBw7USOcEhgNKLyeR5No542lfektPLkXNHBSObuJejP/AVIYMhGUjuU2DiHmLYgEAieagqbTfa0IeV2daGArLRSu24hHI8BU5cdAuDDt+o/4poIBE8fhXl8fJopzesu+5U4AoFAIHiiKXaLozA3sgLb0BuM3EhMJzktCzdn+0ddHYHgqeNZXTFeWtdd5BbHvHnzLM4rVKhg/jx16lS6devGsWPHSq5mzxAJyZlkZBlYv+fio66K4Cln6rJD5q5RgaC4FLnFERkZycCBA/ON69GjB87OzkyYMIG1a9eWWOWedt77Zhd6w73pc7sO32DX4RuolBI/jm6BQvF0L04SlD2idSsoCUpkjKN69eoMGTKE+Pj4kijuqScpJZPlO85R0KJVvUFm6Pd/8c3yw6yKjuXIuQRS0nVlW0nBU4lo3T5blFYLs9AWx+zZs82ftVotc+bMsXIbC2AwGIiNjcXZueCVhk8zU5cdQm2n5P3XXiw03a0kLZv3XubfE7eQZWhc24dsvZH/Tt8mV0NCa5YnqLoXsTdSuHA9ha37rmDIWcnu7aahur8b1XxdqebvSuXyLqhVD6f9YkbX00m23khSaiYJyZkkJmeyZOsZjLJ161aSoFEtH1RKBWqVArVSgUqpQKWU7p3nOXp5pKJNz0KtMqXJjTOdKyzLUUkoi+DOtiz4askBbiSmM/n/GouWVglQZOGQJIlZs2YVmNbT05Mvv/yy5Gr2FHH5Viqb9l7m4OnbqFQKmgf70Ta0Mt7uDsxefQw3Jzvcne2p5udKcnoWTev40rSOye9JVraBy7dSuXAjhQs3kjl79S77YuIAUCklKvu4UKe6N74eDlTzc8XbTfhFeBbQZRtITDGJQkLuMTn3mEFyms5qDbGdSoFOb9qGQgIcNCpcHNXEXk9Gb5DJ1hvRG4xk643ml5WHRSFJqFSSWZCsBUayECardDl51SoF7m4OZGZk5yNo+ZRvIWiSRUurT0SNErm2J4HS6posVDh27twJmFYjvvbaawXOpHJwcMDT07PEKvU0IMsyZ6/eZdO/lzlxMQkHeyXtmlShdUgl3JzszOmGdqtrfuvvExFoVY69WklAJXcCKt1b8XknNYsLN5JNrZIbKWzdexldtmlbd1dHNdX83Kjm50p1P1ee83XFwV4s13nSyNIZcgQhI48gZJKYkklSahZ3Uy33YlIqJDxc7PF201Cnqhdebhq83TR4uZqO7i72/PL7Wf48csPcum1Us3yBD1GjLGPIEZFsg4xebyTbYMTF1YHb8almgTEdZbINBvR62TLccE+I9HqZ7Dxx5ni9kaxsA+kZevP5/enyjgM+LLktLQAnjQqFQkKhkFDm/CkUCtNRklAqc8MkHOzVGAwGU1pJQqlUmPPdn1aZc24Zn6dcRZ54KeeoNLXO8ouPT9ORkpJhXVYBtlQ55wqp9ASz0CeKv/+9bZN79+5tcS64h95g5Or1NJLTsnB1suNobCKb/73M+evJuDiq6d68Gi3rVcRRUzIPcA8XexoElqdBoGnXUU9PJw7H3OLCzRQuXE/mws0UjpxPAExvln7lnKju52oWFD8vJ/PAuxgsfTRkZOktWgiJKSZhyA1Ly8i2SK9SSni5avBy0xBaqwJOdgq83RzMAuHubP/AyRQpWp1V67YgFJKEQqVErVJahJcr54KTqmxbtLki5ubuxK24lAKEySRwFsKkN5Ki1XHwTDxXb6eZrksh4ePhQEAld1RKBUajjMFoamGZPssWn81HWUaXbcyJN1rFF/zZSNG2kS1dcgVTrVQwb0yLhy6vyLvjlgZGo5ElS5YQFRXF9evX8fT05NVXX2X48OE4Ojo+MH+fPn3Yv39/vnErV64kKCjI5joVZ3fcUbP+JjldR83nPEhN13EtPh0vVw1tG1Xmpbq+2KmVDy7kIchvL5v0zGwu5rRIYnO6udIzTVtSa+yUVPV1pZqfK7sOXUebpSe0Znl6tnoBjZ0SezsliiJ2dz0pe/jcz8OO7RRmW5ZltFl6Eu5mmgXBLBA5rYbc3yIXtUph0ULwyvnzdnPAy1WDm7Od+Td5XK+7tCmu7SVbT1u0tFrU87P57fthrtsom4SkIGF5kAi5uGhIuqMtssgZjEa0mdkci03keoLJ+6idSkH9gHL0bPV8kV4QS2x33JiYGCIjIylfvjyjRo0CTDvkTps2jZSUFNq3b8+ECRNQqYr+Vj158mQiIyNp06YN/fv3JzY2lsjISGJiYli8eDGKIgyseXh48PHH1otcKlUqfPfMkmDgN3+Sbbi3dfGpSyaPh0qFxJSBjVEpH93AoJNGTZ1qXtSpZvLUKMsyt+9kEJvTxfXnoeucunzPQ+P+U7fZf+qe4y17tRKNndIsJBo7lfk872dvTyf02XrTuTon3N4yjb1a+VSNu8iyTHJaFhdvpliMLZhEwtR6yMgyWOSxVyvNglC9olsekTC1Glwd1WXyHT2LkyBsaWmVBgpJQqGUoJjvj8UVrYwsA9cTtEiYJks42CtLrFehyE/5qKgoDhw4YHYPe/XqVcaOHYuTkxNhYWFs376dypUr8+677xapvHPnzrF06VLCw8MtBt0rVqzIl19+yaZNm+jYseMDy3F0dKRz585FvYwSZergJvz6x3mLweoGgeXo1eqFRyoa+SFJEj6ejvh4OtK0ji8dmz7H8h3nOHD6NjKgVEpULu9MnapeSJJpUD5Tl/OXpSdTZyA5TUecTm8Oz8o2PNAumLrL7M0iZBIUh0LESGOvshAujZ3KQozsVA/vgKewLjpZlknRZt9rIeQZX0jIaTnosi19HTjYK/FydcDbzYHAyh73hMHdJA5OGtVTJZ5PEkO71X3UVXgklKZgFlk4/vvvP+bNm0e1atUA+O2339Dr9SxYsIDatWtz6tQpxowZU2Th2LhxI7Is8/bbll6pXn/9daZPn8769euLJBxg6vLSarU4OTk99D/ne++9w61bt8znnTp1pX///0Or1fLmmz2s0tdtPRhwylF1A9F/bGPLwrHm+H79BtClS3euX7/GkCHvWeUfPHgYERGvcv78OT74YIRV/KhRY2jevCXHjx9j/PiPrOI/+eQz2rdvzf79+5g8+XOr+EmTviYoqC7R0bv47rtvLOIcn2uNpsKLSEgYDEZOHf6Lg2t2WqSZM2c+/v4VWbt2FYtXLLAqP+q3FWRkqVi9dg2bt25DUtqZ/hSm41t9BmCQFRw8fIQLsVfuxSvVSEo7/PyfI1OnJ1WbhVzEZUWybEQ26FBgwNenHBo7JTevXyHlbgKyQYdsyEY26nB2tOf17j2wt1Oybk0UVy7F5sTrMHgE41GxLvM3xFDrOQ82bN1JWhYo7V1R2LsiKdSWRg1Z6DPvYsxKxpiVAvo0qviVY/C7/fB20zBk8Dtcv5NkkeWll5ozevSHAPTq1Y3MzEyL+DZt2jJkiMmda5cu7ayus6B7T61Wkp1toFevt+jV6y0SExMZMKCPVf7SuPdybYPp3gsNbVSsew/g22//x/PPv8C2bVuYO9d6xqbFvbd4gYVtgAULIvHy8iIqahlRUcus8v/yy0ocHR1ZuPAn1q9fYxW/du3mHDsz+f33rRZxGo2GqKjVAEyfPpW9e/+2sO3h4cmiRUsB+PLLiRw4YNll7uvrx9y5/w+ATz/9kBMnjlvEV6/+PNOnzwRg9OjhxMaet4ivUyeIL7+cCpjGly9evGwRHxISyqefTgTgnXd6c6eAe2/qskOsmjcWP0/ThJwdS0zxD7r3evd+m8GDC36WF1k4UlNTzaIBsG3bNkJCQqhduzYANWvWJDk5uajFceLECRQKBXXrWr4N2NvbU6NGDY4fP15ATkvi4uKoV68emZmZODg40KxZM0aNGkX16tWLXJeHQWdQ4OZkh7O9ntiYwyjsnMrEbkmgsHNEo9DhU84blf4Op5Jsr7uDvQqNnT1OagOG9Dir+NYN/HB0dOTO+Z2cuLjdKn7JpLz/vCbhQalGUtihcXTm0wmTydQZWLNuHWfPX8gRJZPoODq74l+uKpk6PbKkQGHvlkeY7DAoVCzfmePARxOMS41gK/unLt8xddlpnkNhiMOgTUB35wLGrBQqVfBg5JBBeLlpeH/kIJJu3jDnU6uVOPk2oLKPi83fmUDwpFPkwfFWrVqxdetW7OzsOHr0KD179uTzzz+nZ0+TQxSj0UjLli2Jjo4ukuGOHTuSmJjIP//8YxU3YsQItm7dyvHjx7Gzs8snt4mPP/6Y8uXLExgYiEKh4OjRoyxbtgy1Ws0vv/xCYKD19NYHUZzB8aIuACwtxGBp/ugNRlOXms5AZk4XW0JyJn8cusa5a6aXHLVSQfAL3rzZ+gWb+n8f5+sWtoXth6XEBsfr16/PZ599RkREBDNnzsTBwYF27e41cVasWIGvr2+RK5aRkVGgKNjbm/6BMzMzCxWOKVOmWJy3bduWV155hT59+vD111+zaNGiItcnl8K+rIJQ25lGvcqVe3Rvn8W1XRJ1f9Ku+3J8OueuJSNJoDca8fZ05Pmq3mViu6QQtoXtR0mRhWPUqFH079+fNWvWoFQqGT9+PC4uLhiNRjp16kRsbCxjxowpsmEHBwcSExPzjcvKMg3iaDSaIpeXS0hICCEhIezbt4/MzEybyyhOiyNbZ0Btp3wi30hyW0nFzf8kvondTkyzGDSMS0izuZwn8bqFbWG7qJRYi8Pf359NmzZx/vx5PDw88PHxyTGgYMKECQDUqlWryBUrX74858+fR6fTWbUq4uLi8PDwKLS1URgVK1Zk//79JCcnF0t8BE83D1qtLxAICsempcwqlYoaNawXzoSGhtpsuE6dOvz9998cO3aMkJAQc3hWVhanT5+2CLOVS5cuoVKpcHcvGcfsD+LDt+o/0jcSge08i+sZBIKSwubFBmvXrmXgwIG0bduWtm3bMmjQINatW2ez4Xbt2iFJEj///LNF+G+//UZGRobFVNzbt28TGxtLRkaGOSw1NRWDwXodwZ9//smhQ4do2rSpeaxEIBAIBCVHkVscOp2OIUOGsHv3bovwS5cuER0dzcaNG/nhhx9Qq9UFlGBJYGAgb731FkuXLmXo0KE0b97cvHI8NDTUQjhmzJjBmjVrWLJkCY0aNQJg3759TJkyhZYtW1KpUiVUKhXHjh1j/fr1eHh48MknnxT10gQCgUBgA0UWjvnz5xMTE8PYsWNp1aoV5cubNti7ffs2O3fuZOHChcyfP58hQ4YU2fgnn3yCv78/v/76K3/++SceHh707t2b4cOHP3C7kapVq1K7dm3+/PNPEhMTyc7OpkKFCvTq1YtBgwaZx2AEAoFAULIUeR1HREQEM2bMMC/4u5/jx4/zwQcfsG3bthKtYFlTnFlV8OzNuhC2hW1h++m1/aBZVUUe40hPTy9QNACCgoJIT0+3rXYCgUAgeOIosnAolcoC110AJCQkFGk3W4FAIBA82RT5Sf/SSy8xbNgwYmJirOJOnjzJ8OHDad68eYlWTiAQCASPH0UeHB85ciSvvfYa3bt3x8vLy2JwPDExEV9fX2bOnFlqFS0rHuRFrbTyPizCtrAtbAvbZWXPJg+ASUlJTJ8+ne3bt5OaahqscXFxISIigvfff1/4HRcIBIJngGK5jpVlmaSkJGRZxsvLSzioEQgEgmeIBwrHzZs3OXbsmNl3hlgfIRAIBM82hY5xfP3110RGRmI0mtxkKpVK/u///o8RI6y9hQkEAoHg2aBA4Vi+fDmLFy/G19eXWrVqYTQaOXHiBD/++CNVq1alU6dOZVlPgUAgEDwmFNhV1blzZ1588UUmTJiASmXSF51Ox6effsrVq1dZvnx5mVZUIBAIBI8HBQpHcHAwe/bswcnJ0g91QkICHTp0YO/evWVSQYFAIBA8XhS4ANDR0dFKNAC8vb1RKpX55inO9uoCgUAgeLIoUDgK2z6koOm333zzzcPXSCAQCASPNQUOjmdmZrJ27dp847KysvKNy/UVLhAIBIKnlwLHOGrUqFFgy0KWZau43LBTp06VfC0FAoFA8NhQYIvD2dmZcePGFbkgWZaZMmVKiVTqSWDevHmcPHmSkydPcu3aNfz9/fnjjz/KxPaFCxeYM2cOMTEx3L59G71ej6+vL82bN2fAgAHmfcRKi8DAwHzDHR0dOXz4cKnZnTVrFrNnzy4wXqVScfLkyVKzn5CQwMyZM4mOjiYxMRFvb29at27N8OHDcXV1LREbttxXuR4vT5w4wZkzZ9BqtUyZMoVu3bqVuu2FCxeya9cuLl68yN27d3F3d6dq1ar07duXNm3alKrtwu6DsWPHMmDAgFKzXdC9n8vIkSMZPHhwqdgG2LJlCz///DOnT59GkiRq1qzJwIEDy3yD2QKFQ6PR0LVrV5sKmzFjxkNX6ElhxowZuLu7U6tWLfO+XWVFXFwc8fHxtGnTBh8fH1QqFWfPnuW3335j06ZNrFu3Di8vr1KtQ0hICK+//rpFWFHdBheXNm3aULlyZavwM2fOsGDBAlq2bFlqthMTE3n99de5ffs2PXv25IUXXuDcuXNERUVx4MABli9fjoODw0PbseW+io6OZtmyZVSrVo3AwMCHFm1bbB87dgx/f39efvllPDw8SE5OZuvWrQwdOpThw4fb5AnUVtu5fPzxx3h4eFiE1alTxya7ttqeNm1avuGzZ8/mypUrNt+DttieP38+06dPp1atWgwfPhxJkli/fj0DBw5k2rRpZbu2Ti6AjIyMgqIKpDh5nlSuXLli/ty+fXu5ZcuWj7A2JjZv3iwHBATI8+fPL1U7AQEB8ocffliqNmxh/PjxckBAgLxr165Ss/Hll1/KAQEB8oYNGyzCN2zYIAcEBMhz5swpETu23Ffx8fFyenq6LMuyvGXLFjkgIEBetWpVmdjOj+zsbLljx45ycHCwrNfrS832zJkz5YCAAPnq1as22SgJ2/lx8+ZNuUaNGnK3bt1KzXZ8fLxcu3ZtuUOHDrJOpzOH63Q6uUOHDnLDhg3l1NRUm+0XlwKnTmk0GptFqDh5nlQqVar0qKtghb+/PwApKSllYk+n0z1yr48ZGRls2rQJHx8fXnrppVKzs2/fPjQaDe3bt7cIb9euHfb29qxevbpE7NhyX3l7e+Po6Fgidm21nR8qlQofHx8yMjLQ6/VlYjstLc1mWyVlO5dVq1ZhNBp57bXXSs324cOHyc7OpmPHjhYte7VaTYcOHUhOTmbHjh022y8uRfbHIXj8yMrKIj09HZ1Ox/nz5/n2228ByqS/c9u2baxfvx6DwYCnpyft2rVj5MiLQ5deAAAUf0lEQVSRuLi4lLrtvGzZsoW0tDT69OlT4PqikkCn02Fvb281KUShUKDRaLh69SpJSUnPnGuBu3fvYjQauXPnDlu2bGH37t00atQIe3v7UrfdqVMn0tPTUSqV1K1bl8GDB5d5X78sy6xevRoHBwc6dOhQanZ0Oh2Q/8t5btjRo0fp0qVLqdUhL0I4nmBWrFjBpEmTzOf+/v588803hISElKrdunXr0rZtW6pUqUJaWhrR0dEsXbqU/fv3ExUVle/C0dJi5cqVSJJE9+7dS9XOCy+8wPbt2zl16hQ1a9Y0h586dYrk5GTAtJP0syYcERER3L17FzC1OMLDw5k4cWKp2nRxcaFnz57Uq1cPV1dXLl68yM8//8zAgQOZPHlysScHFIe9e/dy7do1unXrhrOzc6nZef755832+vbtaxG3b98+wHT/lRVCOJ5gWrduTbVq1dBqtcTExPDHH3+QlJRU6nZXrFhhcd6lSxcCAwP57rvvWLJkiU2zSh6GCxcucPDgQZo0aVLqXYdvv/02O3bsYOTIkXzyySfmwfHJkyejVqvJzs4mIyOjVOvwODJ79myysrKIi4tj69atZGVlkZaWVqoC2q9fP6uw7t2707FjR6ZMmUJERESZvbzk/i+U9otLYGAgYWFh7Ny5k2nTppntrV69mr/++gswrb0rK4rsc1zw+FGhQgWaNm1qnhL69ddf8+233zJv3rwyr8uAAQNQq9VER0eXmc2VK1cCFKtv2VZCQkKYMWMG6enpvPfee7Rs2ZLBgwfTqFEjWrRoAVCqb5yPKw0bNqRZs2Z0796dn376CScnJ958801zK6ys8PDwoFevXqSkpJTqlPC8JCcn8/vvv1OtWrVSb+UDfPfdd4SHh7Nw4ULatWtHu3bt2LJlC5999hlQtvefaHE8RdSoUYNatWrxyy+/MHDgwDK1rVarKV++PHfu3CkTe3q9nnXr1uHu7l6sdQPF4dVXXyU8PJyzZ8+Snp5O1apV8fLyokePHqhUKqpUqVIm9Xic6dKlC5s2bWL79u1lIuh5yZ0cUlb34Pr169HpdPTo0aNM7Lm5uTFr1iwSEhK4dOkSjo6O1KhRg927dwNQrVq1MqkHCOF46sjMzCzztz3A3F3x4osvlom9Xbt2kZCQQN++fbGzsysTm2ByZpZ3jCM+Pp5Tp07RsGHDElnH8aSTu+3Qo7gHL126BJhmm5UFK1euRK1Wl9mAdC7e3t4W15jbyn/55ZfLrA6iq+oJJD4+Pt/wvXv3cu7cuVJ9eBf0Nvf999+j1+tLdRFeXnK7qcrqbS8/jEYjX375JQaDgUGDBj2yepQ1Wq0232nYBoOBZcuWASa3DKWBXq/Pd6HczZs3iYqKwt3dnXr16pWK7bwcP36c06dP07Jly1JfbPugeqxYsYLQ0NAy6S7LRbQ4isnatWu5ceMGAElJSWRnZ/PDDz8A4OfnV6pvIRMnTiQ+Pp7GjRvj5+dHVlYWJ0+eZPPmzTg5OfHRRx+Vmu25c+dy9OhRGjVqhK+vL1qtlujoaPbt28eLL75Inz59Ss12LnFxcezevZu6des+cAuIkiI9PZ3XXnuNNm3aULFiRVJTU9m4cSMnT55k1KhRNG7cuETs2HJfXb9+3ezK4Pz584CpJXbr1i3A5Iwtt/umJG1fvnyZ3r17ExERQdWqVXF3dycuLo6NGzdy8eJFunbtavNDrKi2tVotr7zyinliiJubGxcvXmTFihVotVqmT59u83qy4vwvl9SLiy22v//+ey5fvkzdunVxdnYmJiaGVatW4ePjU+CK9tKiwE0OBYXTp08f9u/fn29caGgokZGRpWZ78+bNrF27ljNnzpCUlIQkSfj5+REWFsaAAQPw8/MrNds7duxg+fLlnD17lrt376JUKqlSpQqvvvoq77zzTpnM3//xxx/57rvvmDRpktW2J6WFTqfjww8/5OjRo8THx+Pg4EBQUBD9+vUr0YWHttxX+/bts5qamZclS5bQqFGjEredlJTErFmzOHjwILdu3SI9PR1nZ2dq1apF165d6dixY4EbpD6sbZ1Ox+eff86xY8e4desWWq0WDw8P6tevz7vvvkvdunVtsmuL7VwyMzNp1qwZTk5O7Nq1q1AXFCVpe/v27SxYsICLFy+SkZGBn58frVu3ZuDAgSW2V1pREcIhEAgEApsQYxwCgUAgsAkhHAKBQCCwCSEcAoFAILAJIRwCgUAgsAkhHAKBQCCwCSEcAoFAILAJIRwCgUAgsAmxclxgJiMjg9atW5OZmUlaWhrOzs5oNBoMBgNqtZpatWrRunVrOnfunO/+UB999BFHjx5l3bp1Zbp/VGlz+vRp3n77bT744IMy37jPVtq3b098fDzJyck4OjqaPQRqtVpcXV2pW7cuo0aNKtMN8Z4EunbtypUrV0hLS7N54eSziGhxCMw4ODiwZ88exo0bB8C4cePYs2cPe/fuZd26dbRo0YLvv/+ezp07Exsba5X/zp07JCcnP7Qrz8eNXCEtK5e8D8OmTZuYNWsWAP3792fPnj3s2bOHQ4cO8fXXX7N//366d+9u3hBQYGLNmjXm+17wYIRwCIqEp6cnb7zxBr/99hupqam88847Vpstzp07l127dpWoH+zHgeDgYP777z8GDBjwqKtSbCRJokmTJvTp0wetVsvSpUsfdZUETzBCOAQ24e/vz8cff0xcXBzfffedRZxCoSiTvaoeBU+LGOZuCX/x4sVHXBPBk4wQDoHNRERE4OLiwvr1683uKlu1akWDBg0IDAw0+0DOyMggLCyMevXqERgYSExMDKNHj6Zp06Y0a9aM+fPnA3DixAneeOMNQkJC6NGjBydOnLCyKcsykZGRdOjQgdDQUBo3bsyQIUM4c+aMOc0///xDWFgYderUoVWrVpw5c4Y+ffrQqFEjWrduzeLFi63KvXbtGqNHj6Zly5aEhYURERHBhAkTzHVYunQpYWFh1KxZM9+df0+cOMHAgQNp2rQpTZo0oVu3bmzYsMEizYABAwgNDSUwMJCVK1cye/Zs2rRpQ0hICAMGDODq1atW5a5YsYLOnTvTrFkzXn75Zfr06UNkZCQ6na6Iv1L+5G5N5+HhYRX3999/89Zbb9GwYUMaNmxIr1692LFjhzm+pH/PZcuW0bFjR5o2bUpYWBgjRoywELT27dtTq1YtatasSVhYGJs3bwbgwIED5t+kUaNGHDp0CDBtgJj73YaGhhIWFsaYMWO4fv26lf09e/bQrVs36tevT3h4OLNnz8ZgMDzEN/uMIQsE97Fq1So5ICBAXrVqVYFp+vTpIwcEBMj//fefVb69e/dapJ05c6YcEBAgDxgwQD5//rxsNBrlRYsWyQEBAfKyZcvkr776Ss7IyJBTU1Pl119/XW7evLmcnZ1tUcb48ePl4OBgedeuXbLRaJTv3r0rDxkyRA4ODpZPnz5tkbZ3795yaGioPGzYMDkxMVE2GAzyjz/+KAcEBMjbt283p9PpdHJ4eLj84YcfylqtVpZlWT5//rz8yiuvyB9++KFFmS1btpR79+5tEbZ37165Tp068sSJE+WMjAzZaDTKa9askWvUqCH/8MMPVmkDAgLkTp06yRs2bJANBoN848YNuVWrVnLHjh2tvv/g4GD5yJEj5nr+73//kwMCAuSrV68W+Jvcb2vmzJlWcbnl7NixwyJ8/fr1cmBgoDx37lw5KytLzsrKkufNmycHBATIK1assEhbUr9n3bp15ejoaFmWZTklJUUeMmSI3KBBA/ns2bPmdB988IFco0YN+datWxb54+Li5NDQUDkjI0OWZVk2GAzyu+++Kzdt2lQ+fPiwLMuyfOvWLfmNN96QmzZtKsfFxZnz7tu3T65Vq5b8/vvvy+np6bJOp5MjIyPltm3b5nv/CqwRLQ5Bscj1QHb79u0i52nRogXVq1dHkiTefPNN1Go1U6dOZfDgwWg0GpydnXnttde4efMmx44dM+c7cOAAv/76K3379qVFixZIkoSbmxuTJk3CYDBYdZkB3L17l8GDB+Pp6YlCoeCdd95BpVJZvEHHxsZy6dIlwsPDzd77qlevzqBBgyhXrlyh12I0GpkwYQKOjo58/PHHaDQaJEmiS5cuvPzyy8yaNYsrV65Y5fP396dDhw4oFAp8fX3p2LEjZ86csWh17Ny5k6pVq5odcqnVaoYPH07t2rVRq9VF/r7zotPp2LVrF9u3b+ezzz7jlVdeMcelp6fzxRdfEBwczKBBg7Czs8POzo733nuPoKAgvv3223xbOg/7e3bv3t3stc7FxYXPPvuMjIwMvvjiC3Pabt26YTQaWbt2rYXttWvXEhERYfa9sWHDBv766y9GjBhhdiLl4+PDF198QUJCgrk1BDB16lQkSWLcuHE4OjqiVqvp3bs3Pj4+xfpun0WEcAiKhdFoBLDJ70JeXwl2dna4u7vj7+9v0W1SoUIFALMzIoAtW7YAEBYWZlGeh4cHlSpV4t9//7WayaXRaCxcvNrZ2eHh4WEhdO7u7iiVSmbNmsWRI0fM4T169GD06NGFXktMTAyXLl2iadOmVlOPW7VqhcFgYNu2bVb57vdO5+vrC1gKsJeXFzExMcyfP99iJtfq1attergtXLiQsLAwGjduTHBwMMOGDaNjx468+eabFun+/vtvUlJSaNasmVUZQUFB3LlzJ9/upof9PVu0aGFRXrly5QgKCmL//v0kJiYC0LhxY/z9/Vm9erVF2jVr1tCtWzerMu+/R55//nkcHR3NfrkTEhI4ceIEtWvXxtPT0yKtmIJbdMQ6DkGxSEhIAHjgm3le7u9Xz32Y5yX3jTp37ARMHucAhg8fjlKptEifmZmJQqEgOTnZwoVnfn34dnZ2FgJToUIFxo8fz9SpU+nZsyeVK1cmIiKCnj17UqlSpUKvJbc1kd/1ly9f3iJNXgq63rz1Gjp0KGfPnmX69OnMmjWLpk2b0rlzZ8LDw1Gpiv4v279/f4YNGwbA2bNnGTRoEDNmzDC3enLJ/X4XLVrE8uXLLcrIzs7G0dExX3fFD/t75n5PeckNu3z5Ml5eXkiSRNeuXZk9ezYHDx6kQYMG5jGNvO5pc8vMz7GXJEncvXsXMI1pQf5+ycvKV/nTgBAOgc3odDpiYmKws7MjKCioyPnya53Y4j1t0aJFFq2IwihquW+88Qbt2rVj8+bNbNy4kZ9++onFixczbdo02rVrV2A+uRD/Z4XFFaVe5cuXJyoqikOHDrFp0yY2btzIn3/+SXBwMD///LPNrlEBAgICmDRpEv3792fOnDm0b9/e6vcYNmwY/fr1K3KZxf09bf3uunTpwpw5c1izZg0NGjRg9erVdO3aNd/8W7ZsKdQbXm75tnooFFgiuqoENrNp0ybS09Pp3LlzmUy/fe655wCTr/H7SUhIMM/ishVZljEYDLi5ufHGG2+wbNkyVq5ciZOTE1OnTi1SnfIb48l9O69SpUqx6mUwGJBlmfr16zN+/Hj++usvevbsyZEjR6xmbNlCWFgYoaGhXLhwgV27dpnDq1atCuT//ep0Ov755x/S0tKKbfd+CrOX33dXqVIlGjZsyObNm0lKSmL79u107tzZIl9h98i1a9c4evSouay8dvKS24oWPBghHAKbuHLlCtOmTcPHx4cRI0aUic3cN//t27dbxc2ePZtFixYVq9z9+/fTqVMni7CgoCAaNWr0wFXiNWvW5LnnnuPff/+1Gjj+448/UCqVhIeHF6te/fr1Y+vWreZze3t7evfuDUBqamqxyswlt+tq4cKF5rCmTZvi5ubGzp07rd74d+zYwahRo0p0C5m2bdsCEB0dbREeHx/P8ePHCQ0Nteh2BOjevTvp6emMGTOG4OBgq7Gewu6Rzz//nPXr1wOm7qigoCBOnjxJUlKSRbqCfH8LrBHCISgSSUlJLFu2jNdffx03NzcWLVpk0/jGw1C/fn3eeust1q9fz6ZNmzAajRgMBlatWsWGDRsYPnx4scs+f/48y5cvN8/hP3XqFPv27aN9+/aF5lMoFEyaNAmtVsuUKVPIyspClmXWrl3LX3/9xbBhw6hcuXKx6/XTTz+Z1x9kZWURFRWFRqOxmA1VHEJDQwkNDeW///4zz3RycnJi4sSJXL16lWnTppnHIw4fPszkyZN5//33S1Q4QkJC6NmzJ6tWrTIPWqempvL555/j4ODAhAkTrPJERETg5OTE33//bTEonkuHDh1o2bIlixYt4p9//gFMraUff/yRmJgYi1X/Y8eORZZlvvrqK7RaLdnZ2Sxbtoxz586V2DU+7UhyYR2OgmeKwjY5VKlU1KxZk/Dw8Hw3OWzVqhXJycmkpaXh5uZGcHAw8+fPp3Xr1iQmJqLVavHw8KB///7Ur1+fESNGkJSUhFKpxM3NjV9//ZWFCxeyceNGkpOTcXZ2platWkRGRgKmbqWoqCiWL19OXFwcjo6OBAYGMnToUOrUqQOYNiMcMGAAycnJGAwGPD09+eabb1AoFIwePdrC3uLFi/H19SUqKorff/+dGzduIMsyrq6udOnShX79+mFnZ8fSpUuZO3euRd7p06fTuHFjAI4fP86sWbM4fvw4sizj5+fH22+/bdGVMmbMGKKjo83XFRwczIIFC3jvvfc4ePCg+Tt75ZVXmDJlCgcOHOC3337j8OHDZGRkoFAoqF27NkOHDqV27dqF/ob5bXI4btw4i/Gaffv20bdvX3P8hg0b8PT05N9//2Xu3LmcOXMGe3t7KlSowDvvvMOrr75qzluSv+eyZcv49ddfSUhIQJIkQkJCGDlyZIEbMI4bN44dO3awe/fufIUsOzubBQsWsGbNGvP116tXj2HDhpm7snLZs2cP06dP59KlS7i7u9OmTRuee+45Jk6ciJubG0FBQSxYsKDQ7/pZRgiHQCAQCGxCdFUJBAKBwCaEcAgEAoHAJoRwCAQCgcAmhHAIBAKBwCaEcAgEAoHAJoRwCAQCgcAmhHAIBAKBwCaEcAgEAoHAJoRwCAQCgcAmhHAIBAKBwCb+P22UdrXm8nLwAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sn.set()\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "plt.rcParams['font.family'] = 'Serif'\n",
    "\n",
    "df_rlace.plot('rank', 'average', yerr='std', ax=ax, label=\"RLACE (ours)\", marker=\"*\")\n",
    "df_inlp.plot('rank', 'average', yerr='std', ax=ax, label=\"INLP\", marker=\"*\")\n",
    "df_pca.plot('rank', 'average', yerr='std', ax=ax, label=\"PCA-based\", marker=\"*\")\n",
    "#v_measures_original_df.plot('num_clusters', 'average', yerr='std', ax=ax, label=\"Original space\", marker=\"*\", linestyle = \"--\")\n",
    "\n",
    "plt.ylabel(\"Post-projection Accuracy\", fontsize=18, fontname=\"Serif\")\n",
    "plt.xlabel(\"Dimensions Removed\", fontsize=18, fontname=\"Serif\")\n",
    "ax.axhline(get_maj(y_test), label=\"Majority\", color = \"black\", linestyle=\"--\")\n",
    "\n",
    "#plt.legend(fontsize=18)\n",
    "#ax.yaxis.grid(color='gray', linestyle=\"-\")\n",
    "#ax.xaxis.grid(color='gray', linestyle='-')\n",
    "plt.yticks(fontsize=18)\n",
    "plt.xticks(range(1,21,2), fontsize=18)\n",
    "plt.subplots_adjust(bottom=0.17)\n",
    "plt.subplots_adjust(left=0.15)\n",
    "ax.figure.savefig(\"analysis/analysis-results/gender-glove.pdf\", dpi = 700) \n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5058503401360545"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## load glove"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# #with open(\"glove-all-gendered_data.pickle\", \"rb\") as f:\n",
    "# #    all_data = pickle.load(f)    \n",
    "# #vecs, words, labels = all_data[\"vecs\"], all_data[\"words\"], all_data[\"labels\"]\n",
    "# model_all, vecs_all, words_all =  load_word_vectors(fname = \"../../glove.42B.300d.txt\")\n",
    "# rank=1\n",
    "\n",
    "# #vecs_cleaned = (P.dot(vecs_all.T)).T\n",
    "# #vecs_cleaned_inlp = (get_inlp_p(ws_inlp, rank)@vecs_all.T).T\n",
    "\n",
    "# \"\"\"\n",
    "# save_in_word2vec_format(vecs_cleaned, words_all, \"../../glove.42B.300d.cleaned.rank={}.txt\".format(rank))\n",
    "# save_in_word2vec_format(vecs_cleaned_inlp, words_all, \"../../glove.42B.300d.cleaned-inlp.rank={}.txt\".format(rank))\n",
    "\n",
    "\n",
    "# model_cleaned, _, _ = load_word_vectors(fname = \"glove.42B.300d.cleaned.rank={}.txt\".format(rank))\n",
    "# model_cleaned_inlp, _, _ = load_word_vectors(fname = \"glove.42B.300d.cleaned-inlp.rank={}.txt\".format(rank))\n",
    "#  \"\"\"\n",
    "\n",
    "# import copy\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "def project_model(model, words, vecs, P):\n",
    "    \n",
    "    new = copy.deepcopy(model)\n",
    "    \n",
    "    vecs_proj = vecs@P\n",
    "    for w,v in tqdm.tqdm_notebook(zip(words, vecs_proj)):\n",
    "        new.__setitem__(w,v)\n",
    "    \n",
    "    return new"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## WEAT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Auxiliary functions for experiments by Caliskan et al.\n",
    "\n",
    "import scipy\n",
    "import scipy.misc as misc\n",
    "import itertools\n",
    "\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "\n",
    "def get_sim(vecs, w2ind, w1, w2):\n",
    "    v1,v2 = vecs[w2ind[w1]], vecs[w2ind[w2]]\n",
    "    sim = cosine_similarity([v1],[v2])[0]\n",
    "    return sim\n",
    "    \n",
    "\n",
    "def s_word(w, A, B, vecs, w2ind, all_s_words):\n",
    "    \n",
    "    if w in all_s_words:\n",
    "        return all_s_words[w]\n",
    "    \n",
    "    mean_a = []\n",
    "    mean_b = []\n",
    "    \n",
    "    for a in A:\n",
    "        mean_a.append(get_sim(vecs, w2ind, w, a))# model.similarity(w,a))\n",
    "    for b in B:\n",
    "        mean_b.append(get_sim(vecs, w2ind, w, b)) # model.similarity(w,b))\n",
    "        \n",
    "    mean_a = sum(mean_a)/float(len(mean_a))\n",
    "    mean_b = sum(mean_b)/float(len(mean_b))\n",
    "    \n",
    "    all_s_words[w] = mean_a - mean_b\n",
    "\n",
    "    return all_s_words[w]\n",
    "\n",
    "\n",
    "def s_group(X, Y, A, B, vecs, w2ind, all_s_words):\n",
    "    \n",
    "    total = 0\n",
    "    total_x, total_y = 0.,0.\n",
    "    all_vals = []\n",
    "    \n",
    "    for x in X:\n",
    "        x_sim = s_word(x, A, B, vecs, w2ind, all_s_words)\n",
    "        total += x_sim\n",
    "        total_x += x_sim\n",
    "        all_vals.append(x_sim)\n",
    "    for y in Y:\n",
    "        y_sim =  s_word(y, A, B, vecs, w2ind, all_s_words)\n",
    "        total -= y_sim\n",
    "        total_y += y_sim\n",
    "        all_vals.append(y_sim)\n",
    "    \n",
    "    #print(x_sim, y_sim)\n",
    "    numerator = total_x/len(X) - total_y/len(Y)\n",
    "    \n",
    "    return total,numerator/np.std(all_vals)\n",
    "\n",
    "\n",
    "def p_value_exhust(X, Y, A, B, vecs, w2ind):\n",
    "    [\"Ps\"]\n",
    "    if len(X) > 10:\n",
    "        print ('might take too long, use sampled version: p_value')\n",
    "        return\n",
    "    \n",
    "    assert(len(X) == len(Y))\n",
    "    \n",
    "    all_s_words = {}\n",
    "    s_orig,d_statistic = s_group(X, Y, A, B, vecs, w2ind, all_s_words)\n",
    "    #print(\"s-orig: {}\".format(s_orig))\n",
    "    \n",
    "    union = set(X+Y)\n",
    "    subset_size = int(len(union)/2)\n",
    "    \n",
    "    larger = 0\n",
    "    total = 0\n",
    "    #all_subs = set(itertools.combinations(union, subset_size))\n",
    "    #print(all_subs)\n",
    "    s_group_vals = []\n",
    "    \n",
    "    for subset in set(itertools.combinations(union, subset_size)):\n",
    "        total += 1\n",
    "        Xi = list(set(subset))\n",
    "        Yi = list(union - set(subset))\n",
    "        \n",
    "        \n",
    "        val,_ = s_group(Xi, Yi, A, B, vecs, w2ind, all_s_words)\n",
    "        s_group_vals.append(val)\n",
    "        \n",
    "        if val > s_orig:\n",
    "            larger += 1\n",
    "    #print ('num of samples', total)\n",
    "    \n",
    "    return larger/float(total), d_statistic\n",
    "\n",
    "\n",
    "def p_value_sample(X, Y, A, B, vecs, w2ind):\n",
    "    \n",
    "    random.seed(10)\n",
    "    np.random.seed(10)\n",
    "    all_s_words = {}\n",
    "    \n",
    "    assert(len(X) == len(Y))\n",
    "    length = len(X)\n",
    "    \n",
    "    s_orig,_ = s_group(X, Y, A, B, vecs, w2ind, all_s_words) \n",
    "    \n",
    "    num_of_samples = min(10000, int(scipy.special.comb(length*2,length)*100))\n",
    "    print ('num of samples', num_of_samples)\n",
    "    larger = 0\n",
    "    for i in range(num_of_samples):\n",
    "        permute = np.random.permutation(X+Y)\n",
    "        Xi = permute[:length]\n",
    "        Yi = permute[length:]\n",
    "        if s_group(Xi, Yi, A, B, vecs, w2ind, all_s_words) > s_orig:\n",
    "            larger += 1\n",
    "    \n",
    "    return larger/float(num_of_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "w2ind = {w:i for i,w in enumerate(words_all)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Relaxed (0.07420357420357421, array([0.77301392]))\n",
      "INLP (0.011732711732711733, array([1.08871693]))\n",
      "PCA (0.7153846153846154, array([-0.40178972], dtype=float32))\n",
      "original (7.77000777000777e-05, array([1.694486], dtype=float32))\n",
      "======================================================\n",
      "Relaxed (0.06045066045066045, array([0.80773742]))\n",
      "INLP (0.03325563325563326, array([0.94080575]))\n",
      "PCA (0.038461538461538464, array([0.7184212], dtype=float32))\n",
      "original (7.77000777000777e-05, array([1.5671507], dtype=float32))\n",
      "======================================================\n",
      "Relaxed (0.06822066822066822, array([0.7797685]))\n",
      "INLP (0.05485625485625486, array([0.8194422]))\n",
      "PCA (0.19324009324009325, array([0.30151135], dtype=float32))\n",
      "original (0.0001554001554001554, array([1.6302873], dtype=float32))\n"
     ]
    }
   ],
   "source": [
    "# Experiment 1\n",
    "vecs_cleaned_rlace = vecs_all @ projs_rlace[0][1]\n",
    "vecs_cleaned_inlp = vecs_all @ projs_inlp[0][0]\n",
    "vecs_cleaned_pca = vecs_all @ projs_words[0]\n",
    "vecs_original = vecs_all\n",
    "\n",
    "A = ['john', 'paul', 'mike', 'kevin', 'steve', 'greg', 'jeff', 'bill']\n",
    "B = ['amy', 'joan', 'lisa', 'sarah', 'diana', 'kate', 'ann', 'donna']\n",
    "\n",
    "C = ['executive', 'management', 'professional', 'corporation', 'salary', 'office', 'business', 'career']\n",
    "D = ['home', 'parents', 'children', 'family', 'cousins', 'marriage', 'wedding', 'relatives']\n",
    "\n",
    "print (\"Relaxed\", p_value_exhust(A, B, C, D, vecs_cleaned_rlace, w2ind))\n",
    "print (\"INLP\", p_value_exhust(A, B, C, D, vecs_cleaned_inlp, w2ind))\n",
    "print (\"PCA\", p_value_exhust(A, B, C, D, vecs_cleaned_pca, w2ind))\n",
    "print (\"original\", p_value_exhust(A, B, C, D, vecs_original, w2ind))\n",
    "\n",
    "print(\"======================================================\")\n",
    "\n",
    "# Experiment 2\n",
    "\n",
    "E = ['math', 'algebra', 'geometry', 'calculus', 'equations', 'computation', 'numbers', 'addition']\n",
    "F = ['poetry', 'art', 'dance', 'literature', 'novel', 'symphony', 'drama', 'sculpture']\n",
    "\n",
    "print (\"Relaxed\", p_value_exhust(A, B, E, F, vecs_cleaned_rlace, w2ind))\n",
    "print (\"INLP\", p_value_exhust(A, B, E, F, vecs_cleaned_inlp, w2ind))\n",
    "print (\"PCA\", p_value_exhust(A, B, E, F, vecs_cleaned_pca, w2ind))\n",
    "print (\"original\", p_value_exhust(A, B, E, F, vecs_original, w2ind))\n",
    "\n",
    "print(\"======================================================\")\n",
    "\n",
    "# Experiment 3\n",
    "\n",
    "G = ['science', 'technology', 'physics', 'chemistry', 'einstein', 'nasa', 'experiment', 'astronomy']\n",
    "H = ['poetry', 'art', 'shakespeare', 'dance', 'literature', 'novel', 'symphony', 'drama']\n",
    "\n",
    "print (\"Relaxed\", p_value_exhust(A, B, G, H, vecs_cleaned_rlace, w2ind))\n",
    "print (\"INLP\", p_value_exhust(A, B, G, H, vecs_cleaned_inlp, w2ind))\n",
    "print (\"PCA\", p_value_exhust(A, B, G, H, vecs_cleaned_pca, w2ind))\n",
    "print (\"original\", p_value_exhust(A, B, G, H, vecs_original, w2ind))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ecc0505a457549b7b278da3c8e09f58e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "502e26d4df9343bab7b36fff06671f05",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3e1ac92c27544e9a97964a71ead90a96",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a3558616bd39491aab2c25a33db44538",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b754db94699a4a0fb193588ae5004d77",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "59b81b778c5d4224a90e05837c296261",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6e1f3cc1e51148738df30d65dd405cc2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "83928cafd99e402bb2ecd45c7cbf9ae9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1f92c4d2d86d49e4af14780f68766921",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "vecs_original = vecs_all  \n",
    "weat = defaultdict(list)\n",
    "weat_type2pair = {\"profession-family\": (C,D), \"math-art\": (E,F), \"science-art\": (G,H)}\n",
    "\n",
    "for weat_type in [\"profession-family\", \"math-art\", \"science-art\"]:\n",
    "    for method in [\"PCA\", \"RLACE\", \"INLP\"]:\n",
    "    \n",
    "        ps = []\n",
    "        ds = []\n",
    "        \n",
    "        for seed in tqdm.tqdm_notebook(range(5)):\n",
    "                \n",
    "                vecs_cleaned_rlace = vecs_all @ projs_rlace[seed][1]\n",
    "                vecs_cleaned_inlp = vecs_all @ projs_inlp[seed][0]\n",
    "                vecs_cleaned_pca = vecs_all @ projs_words[0]\n",
    "                method2vecs = {\"PCA\": vecs_cleaned_pca, \"INLP\": vecs_cleaned_inlp, \"RLACE\": vecs_cleaned_rlace}\n",
    "\n",
    "                p,d = p_value_exhust(A, B, *weat_type2pair[weat_type], method2vecs[method], w2ind)\n",
    "                d = d.item()\n",
    "                ps.append(p)\n",
    "                ds.append(d)\n",
    "        mean_p, std_p = np.mean(ps), np.std(ps)\n",
    "        mean_d, std_d = np.mean(ds), np.std(ds)\n",
    "        d_str = \"{:.2f} $\\pm$ {:.2f}\".format(mean_d, std_d)\n",
    "        p_str = \"{:.3f} $\\pm$ {:.3f}\".format(mean_p, std_p)\n",
    "        weat[weat_type].append({\"Method\": method, \"p\": p_str, \"d\": d_str})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "p_orig,d_orig = p_value_exhust(A, B, C,D, vecs_all, w2ind)\n",
    "d_orig=d_orig.item()\n",
    "\n",
    "p_orig2,d_orig2 = p_value_exhust(A, B, E,F, vecs_all, w2ind)\n",
    "d_orig2=d_orig2.item()\n",
    "\n",
    "p_orig3,d_orig3 = p_value_exhust(A, B, G,H, vecs_all, w2ind)\n",
    "d_orig3=d_orig3.item()\n",
    "\n",
    "\n",
    "d_str = \"{:.2f}\".format(d_orig)\n",
    "p_str = \"{:.3f}\".format(p_orig)\n",
    "d_str2 = \"{:.2f}\".format(d_orig2)\n",
    "p_str2 = \"{:.3f}\".format(p_orig2)\n",
    "d_str3 = \"{:.2f}\".format(d_orig3)\n",
    "p_str3 = \"{:.3f}\".format(p_orig3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "weat[\"profession-family\"] = [{\"Method\": \"Original\", \"d\": d_str, \"p\": p_str}] + weat[\"profession-family\"]\n",
    "weat[\"math-art\"] = [{\"Method\": \"Original\", \"d\": d_str2, \"p\": p_str2}] + weat[\"math-art\"] \n",
    "weat[\"science-art\"] = [{\"Method\": \"Original\", \"d\": d_str3, \"p\": p_str3}] + weat[\"science-art\"] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Method</th>\n",
       "      <th>d</th>\n",
       "      <th>p</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Original</td>\n",
       "      <td>1.57</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PCA</td>\n",
       "      <td>0.72 $\\pm$ 0.00</td>\n",
       "      <td>0.038 $\\pm$ 0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RLACE</td>\n",
       "      <td>0.80 $\\pm$ 0.01</td>\n",
       "      <td>0.062 $\\pm$ 0.002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>INLP</td>\n",
       "      <td>1.10 $\\pm$ 0.10</td>\n",
       "      <td>0.016 $\\pm$ 0.009</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Method                d                  p\n",
       "0  Original             1.57              0.000\n",
       "1       PCA  0.72 $\\pm$ 0.00  0.038 $\\pm$ 0.000\n",
       "2     RLACE  0.80 $\\pm$ 0.01  0.062 $\\pm$ 0.002\n",
       "3      INLP  1.10 $\\pm$ 0.10  0.016 $\\pm$ 0.009"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(weat[\"math-art\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Method</th>\n",
       "      <th>d</th>\n",
       "      <th>p</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Original</td>\n",
       "      <td>1.63</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PCA</td>\n",
       "      <td>0.30 $\\pm$ 0.00</td>\n",
       "      <td>0.193 $\\pm$ 0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RLACE</td>\n",
       "      <td>0.77 $\\pm$ 0.01</td>\n",
       "      <td>0.073 $\\pm$ 0.003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>INLP</td>\n",
       "      <td>1.03 $\\pm$ 0.11</td>\n",
       "      <td>0.022 $\\pm$ 0.016</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Method                d                  p\n",
       "0  Original             1.63              0.000\n",
       "1       PCA  0.30 $\\pm$ 0.00  0.193 $\\pm$ 0.000\n",
       "2     RLACE  0.77 $\\pm$ 0.01  0.073 $\\pm$ 0.003\n",
       "3      INLP  1.03 $\\pm$ 0.11  0.022 $\\pm$ 0.016"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(weat[\"science-art\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Method</th>\n",
       "      <th>d</th>\n",
       "      <th>p</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Original</td>\n",
       "      <td>1.69</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PCA</td>\n",
       "      <td>-0.40 $\\pm$ 0.00</td>\n",
       "      <td>0.715 $\\pm$ 0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RLACE</td>\n",
       "      <td>0.78 $\\pm$ 0.01</td>\n",
       "      <td>0.072 $\\pm$ 0.003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>INLP</td>\n",
       "      <td>1.15 $\\pm$ 0.07</td>\n",
       "      <td>0.007 $\\pm$ 0.003</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Method                 d                  p\n",
       "0  Original              1.69              0.000\n",
       "1       PCA  -0.40 $\\pm$ 0.00  0.715 $\\pm$ 0.000\n",
       "2     RLACE   0.78 $\\pm$ 0.01  0.072 $\\pm$ 0.003\n",
       "3      INLP   1.15 $\\pm$ 0.07  0.007 $\\pm$ 0.003"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(weat[\"profession-family\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "with pd.option_context(\"max_colwidth\", 1000):\n",
    "    \n",
    "    pd.DataFrame(weat[\"math-art\"]).to_latex(buf=\"table-weat-math-art-rlace.latex\", encoding='utf-8',escape=False,index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "with pd.option_context(\"max_colwidth\", 1000):\n",
    "    \n",
    "    pd.DataFrame(weat[\"profession-family\"]).to_latex(buf=\"table-weat-profession-family-rlace.latex\", encoding='utf-8',escape=False,index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "with pd.option_context(\"max_colwidth\", 1000):\n",
    "    \n",
    "    pd.DataFrame(weat[\"science-art\"]).to_latex(buf=\"table-weat-science-art-rlace.latex\", encoding='utf-8',escape=False,index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Smantic content"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlp/ravfogs/anaconda3/envs/py3/lib/python3.7/site-packages/ipykernel_launcher.py:40: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5a1e92a4e3b1464c93c39606a037ab6f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=5.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "2\n",
      "4\n",
      "8\n",
      "12\n",
      "16\n",
      "20\n",
      "1\n",
      "2\n",
      "4\n",
      "8\n",
      "12\n",
      "16\n",
      "20\n",
      "1\n",
      "2\n",
      "4\n",
      "8\n",
      "12\n",
      "16\n",
      "20\n",
      "1\n",
      "2\n",
      "4\n",
      "8\n",
      "12\n",
      "16\n",
      "20\n",
      "1\n",
      "2\n",
      "4\n",
      "8\n",
      "12\n",
      "16\n",
      "20\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import pandas\n",
    "\n",
    "\n",
    "def get_corr(vecs, w2ind):\n",
    "    \n",
    "    \n",
    "    df = pandas.read_csv(\"glove_data/SimLex-999.txt\", sep=\"\\t\")\n",
    "    word1,word2,sim = df[\"word1\"].tolist(), df[\"word2\"].tolist(), df[\"SimLex999\"].tolist()\n",
    "    sim_pred = []\n",
    "    sim = np.array(sim)\n",
    "\n",
    "    has_w1 = [True if w in w2ind else False for w in word1]\n",
    "    has_w2 = [True if w in w2ind else False for w in word2]\n",
    "    has_w = np.array([True if (p and q) else False for p,q in zip(has_w1,has_w2)])\n",
    "\n",
    "\n",
    "    ignored = []\n",
    "    for w1,w2,s in zip(word1, word2, sim):\n",
    "        if w1 not in w2ind or w2 not in w2ind: \n",
    "            #print(\"Skipping pair {}, {}\".format(w1, w2))\n",
    "            ignored.append(w1)\n",
    "            ignored.append(w2)\n",
    "            continue\n",
    "        sim_pred.append(get_sim(vecs, w2ind, w1, w2))\n",
    "        #sim_pred_relaxed.append(get_sim(vecs_cleaned_rlace, w2ind, w1, w2))\n",
    "        #sim_pred_inlp.append(get_sim(vecs_cleaned_inlp, w2ind, w1, w2))\n",
    "\n",
    "    sim = sim[has_w]\n",
    "    return scipy.stats.pearsonr(sim,sim_pred)[0]\n",
    "    #print(\"Correlation original: {}; Relaxed: {}; INLP: {}\".format(scipy.stats.pearsonr(sim,sim_pred_orig), \n",
    "    #                                                           scipy.stats.pearsonr(sim,sim_pred_relaxed), \n",
    "    #                                                           scipy.stats.pearsonr(sim,sim_pred_inlp)))\n",
    "    \n",
    "    \n",
    "originals = defaultdict(dict)\n",
    "rlace = defaultdict(dict)\n",
    "inlp = defaultdict(dict)\n",
    "pca = defaultdict(dict)\n",
    "\n",
    "for rand_seed in tqdm.tqdm_notebook(range(5)):    \n",
    "    vecs_cleaned_rlace = vecs_all @ projs_rlace[rand_seed][1]\n",
    "    vecs_cleaned_inlp = vecs_all @ projs_inlp[rand_seed][0]\n",
    "    vecs_cleaned_pca = vecs_all @ projs_words[0]\n",
    "\n",
    "    originals[0][0] = get_corr(vecs_all, w2ind)[0]\n",
    "    \n",
    "    for i,rank in enumerate(projs_rlace[rand_seed].keys()):\n",
    "        vecs_cleaned_rlace = vecs_all @ projs_rlace[rand_seed][rank]\n",
    "        print(rank)\n",
    "        vecs_cleaned_inlp = vecs_all @ (projs_inlp[rand_seed][rank-1] )\n",
    "        rlace[rand_seed][rank] = get_corr(vecs_cleaned_rlace, w2ind)[0]\n",
    "        inlp[rand_seed][rank] = get_corr(vecs_cleaned_inlp, w2ind)[0]\n",
    "        \n",
    "    for rank in range(len(projs_words)):\n",
    "        vecs_cleaned_pca = vecs_all @ projs_words[rank]\n",
    "        pca[0][rank+1] = get_corr(vecs_cleaned_pca, w2ind)[0]\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for k in originals:\n",
    "#     for k2 in originals[k]:\n",
    "#         originals[k][k2] = originals[k][k2].item()\n",
    "# for k in rlace:\n",
    "#     for k2 in rlace[k]:\n",
    "#         rlace[k][k2] = rlace[k][k2].item()\n",
    "# for k in inlp:\n",
    "#     for k2 in inlp[k]:\n",
    "#         inlp[k][k2] = inlp[k][k2].item()\n",
    "# for k in pca:\n",
    "#     for k2 in pca[k]:\n",
    "#         pca[k][k2] = pca[k][k2].item()\n",
    "        \n",
    "df_originals = pd.DataFrame(originals)\n",
    "df_rlace = pd.DataFrame(rlace)\n",
    "df_inlp = pd.DataFrame(inlp)\n",
    "df_pca = pd.DataFrame(pca)\n",
    "\n",
    "for df in [df_originals, df_rlace, df_inlp, df_pca]:\n",
    "    \n",
    "    df['average'] = df.mean(numeric_only=True, axis=1)\n",
    "    df[\"std\"] = df.std(numeric_only=True, axis=1)\n",
    "    df.rename_axis(\"Rank\", inplace=True)\n",
    "    df.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Rank</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.394646</td>\n",
       "      <td>0.395354</td>\n",
       "      <td>0.395164</td>\n",
       "      <td>0.395119</td>\n",
       "      <td>0.395427</td>\n",
       "      <td>0.395142</td>\n",
       "      <td>0.000273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0.394551</td>\n",
       "      <td>0.396678</td>\n",
       "      <td>0.395174</td>\n",
       "      <td>0.394141</td>\n",
       "      <td>0.394080</td>\n",
       "      <td>0.394925</td>\n",
       "      <td>0.000959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>0.392764</td>\n",
       "      <td>0.397447</td>\n",
       "      <td>0.395771</td>\n",
       "      <td>0.392791</td>\n",
       "      <td>0.392831</td>\n",
       "      <td>0.394321</td>\n",
       "      <td>0.001942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>0.391651</td>\n",
       "      <td>0.394828</td>\n",
       "      <td>0.395624</td>\n",
       "      <td>0.390421</td>\n",
       "      <td>0.389623</td>\n",
       "      <td>0.392429</td>\n",
       "      <td>0.002387</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12</td>\n",
       "      <td>0.386403</td>\n",
       "      <td>0.393794</td>\n",
       "      <td>0.393503</td>\n",
       "      <td>0.388746</td>\n",
       "      <td>0.388302</td>\n",
       "      <td>0.390150</td>\n",
       "      <td>0.002965</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>16</td>\n",
       "      <td>0.389442</td>\n",
       "      <td>0.390877</td>\n",
       "      <td>0.391766</td>\n",
       "      <td>0.388672</td>\n",
       "      <td>0.390344</td>\n",
       "      <td>0.390220</td>\n",
       "      <td>0.001081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>20</td>\n",
       "      <td>0.396127</td>\n",
       "      <td>0.389760</td>\n",
       "      <td>0.393424</td>\n",
       "      <td>0.392856</td>\n",
       "      <td>0.393180</td>\n",
       "      <td>0.393070</td>\n",
       "      <td>0.002024</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Rank         0         1         2         3         4   average       std\n",
       "0     1  0.394646  0.395354  0.395164  0.395119  0.395427  0.395142  0.000273\n",
       "1     2  0.394551  0.396678  0.395174  0.394141  0.394080  0.394925  0.000959\n",
       "2     4  0.392764  0.397447  0.395771  0.392791  0.392831  0.394321  0.001942\n",
       "3     8  0.391651  0.394828  0.395624  0.390421  0.389623  0.392429  0.002387\n",
       "4    12  0.386403  0.393794  0.393503  0.388746  0.388302  0.390150  0.002965\n",
       "5    16  0.389442  0.390877  0.391766  0.388672  0.390344  0.390220  0.001081\n",
       "6    20  0.396127  0.389760  0.393424  0.392856  0.393180  0.393070  0.002024"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_inlp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Rank</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.392260</td>\n",
       "      <td>0.392466</td>\n",
       "      <td>0.392354</td>\n",
       "      <td>0.392393</td>\n",
       "      <td>0.392290</td>\n",
       "      <td>0.392353</td>\n",
       "      <td>0.000074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0.392309</td>\n",
       "      <td>0.392468</td>\n",
       "      <td>0.392738</td>\n",
       "      <td>0.393184</td>\n",
       "      <td>0.391761</td>\n",
       "      <td>0.392492</td>\n",
       "      <td>0.000471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4</td>\n",
       "      <td>0.389404</td>\n",
       "      <td>0.391196</td>\n",
       "      <td>0.391726</td>\n",
       "      <td>0.391849</td>\n",
       "      <td>0.393374</td>\n",
       "      <td>0.391510</td>\n",
       "      <td>0.001278</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>0.390623</td>\n",
       "      <td>0.392150</td>\n",
       "      <td>0.392385</td>\n",
       "      <td>0.392497</td>\n",
       "      <td>0.391311</td>\n",
       "      <td>0.391793</td>\n",
       "      <td>0.000718</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12</td>\n",
       "      <td>0.391522</td>\n",
       "      <td>0.394520</td>\n",
       "      <td>0.390108</td>\n",
       "      <td>0.389775</td>\n",
       "      <td>0.388597</td>\n",
       "      <td>0.390904</td>\n",
       "      <td>0.002034</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>16</td>\n",
       "      <td>0.396603</td>\n",
       "      <td>0.391596</td>\n",
       "      <td>0.387501</td>\n",
       "      <td>0.393288</td>\n",
       "      <td>0.391805</td>\n",
       "      <td>0.392159</td>\n",
       "      <td>0.002938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>20</td>\n",
       "      <td>0.390144</td>\n",
       "      <td>0.385910</td>\n",
       "      <td>0.390436</td>\n",
       "      <td>0.390817</td>\n",
       "      <td>0.391207</td>\n",
       "      <td>0.389703</td>\n",
       "      <td>0.001930</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Rank         0         1         2         3         4   average       std\n",
       "0     1  0.392260  0.392466  0.392354  0.392393  0.392290  0.392353  0.000074\n",
       "1     2  0.392309  0.392468  0.392738  0.393184  0.391761  0.392492  0.000471\n",
       "2     4  0.389404  0.391196  0.391726  0.391849  0.393374  0.391510  0.001278\n",
       "3     8  0.390623  0.392150  0.392385  0.392497  0.391311  0.391793  0.000718\n",
       "4    12  0.391522  0.394520  0.390108  0.389775  0.388597  0.390904  0.002034\n",
       "5    16  0.396603  0.391596  0.387501  0.393288  0.391805  0.392159  0.002938\n",
       "6    20  0.390144  0.385910  0.390436  0.390817  0.391207  0.389703  0.001930"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_rlace"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Rank</th>\n",
       "      <th>0</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.168936</td>\n",
       "      <td>0.168936</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0.340804</td>\n",
       "      <td>0.340804</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>0.319348</td>\n",
       "      <td>0.319348</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0.384118</td>\n",
       "      <td>0.384118</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>0.395134</td>\n",
       "      <td>0.395134</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>0.386605</td>\n",
       "      <td>0.386605</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>0.396802</td>\n",
       "      <td>0.396802</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>0.387988</td>\n",
       "      <td>0.387988</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>0.404865</td>\n",
       "      <td>0.404865</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>0.409585</td>\n",
       "      <td>0.409585</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Rank         0   average  std\n",
       "0     1  0.168936  0.168936  0.0\n",
       "1     2  0.340804  0.340804  0.0\n",
       "2     3  0.319348  0.319348  0.0\n",
       "3     4  0.384118  0.384118  0.0\n",
       "4     5  0.395134  0.395134  0.0\n",
       "5     6  0.386605  0.386605  0.0\n",
       "6     7  0.396802  0.396802  0.0\n",
       "7     8  0.387988  0.387988  0.0\n",
       "8     9  0.404865  0.404865  0.0\n",
       "9    10  0.409585  0.409585  0.0"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_pca"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Rank</th>\n",
       "      <th>0</th>\n",
       "      <th>average</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.399052</td>\n",
       "      <td>0.399052</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Rank         0   average  std\n",
       "0     0  0.399052  0.399052  0.0"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_originals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [],
   "source": [
    "common_vecs, common_ws = vecs_all[:15000], words_all[:15000]\n",
    "common_vecs_transformed =  common_vecs@projs_rlace[0][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "history ['literature', 'histories', 'historical'] ['literature', 'histories', 'historical']\n",
      "1989 ['1991', '1987', '1988'] ['1991', '1987', '1988']\n",
      "afternoon ['saturday', 'evening', 'morning'] ['sunday', 'evening', 'morning']\n",
      "consumers ['buyers', 'customers', 'consumer'] ['buyers', 'customers', 'consumer']\n",
      "allowed ['allowing', 'allow', 'permitted'] ['allowing', 'allow', 'permitted']\n",
      "leg ['thigh', 'knee', 'legs'] ['thigh', 'knee', 'legs']\n",
      "manner ['therefore', 'regard', 'thus'] ['means', 'regard', 'thus']\n",
      "vinyl ['metal', 'lp', 'pvc'] ['metal', 'lp', 'pvc']\n",
      "injury ['injured', 'accident', 'injuries'] ['injured', 'accident', 'injuries']\n",
      "worried ['afraid', 'concerned', 'worry'] ['afraid', 'concerned', 'worry']\n",
      "dishes ['cooking', 'cuisine', 'dish'] ['meals', 'cuisine', 'dish']\n",
      "thursday ['monday', 'tuesday', 'wednesday'] ['monday', 'tuesday', 'wednesday']\n",
      "sisters ['brothers', 'daughters', 'sister'] ['daughters', 'brothers', 'sister']\n",
      "wants ['decides', 'thinks', 'knows'] ['decides', 'thinks', 'knows']\n",
      "covering ['cover', 'covered', 'covers'] ['cover', 'covered', 'covers']\n"
     ]
    }
   ],
   "source": [
    "sims_original = cosine_similarity(common_vecs,common_vecs)\n",
    "sims_transformed = cosine_similarity(common_vecs_transformed,common_vecs_transformed)\n",
    "sims_transformed = sims_transformed.argsort(axis=1)\n",
    "sims_original = sims_original.argsort(axis=1)\n",
    "\n",
    "words = []\n",
    "\n",
    "for i in range(15):\n",
    "    j = np.random.choice(range(5000))\n",
    "    w = common_ws[j]\n",
    "    before = [common_ws[k] for k in sims_original[j][-4:-1]]\n",
    "    after = [common_ws[k] for k in sims_transformed[j][-4:-1]]\n",
    "    print(w, before, after)\n",
    "    words.append({\"Word\": w, \"Neighbors before\": \", \".join(before), \"Neighbors after\": \", \".join(after)})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(words)\n",
    "with pd.option_context(\"max_colwidth\", 1000):\n",
    "    \n",
    "    df.to_latex(buf=\"table-neighbors-rlace.latex\", encoding='utf-8',escape=False,index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lawyer ['malpractice', 'litigation', 'attorneys', 'lawyers', 'attorney'] ['litigation', 'malpractice', 'attorneys', 'lawyers', 'attorney']\n"
     ]
    }
   ],
   "source": [
    "\n",
    "words = []\n",
    "\n",
    "word_set = [\"lucy\", \"john\", \"mary\", \"michael\", \"edith\", \"jonathan\", \"daniel\", \"ann\", \"anna\", \"olivia\", \"james\", \"sophia\"]\n",
    "\n",
    "female_names = [\"lucy\", \"mary\", \"edith\", \"ann\", \"olivia\", \"ann\", \"alice\", \"sophia\", \"emma\", \"naomi\", \"claire\", \"gabriella\",\n",
    "               \"clara\", \"jennifer\", \"judy\", \"angelina\", \"helen\"]\n",
    "male_names = [\"john\", \"michael\", \"jonathan\", \"daniel\", \"james\", \"samuel\"]\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "for w in [\"lawyer\"]: #female_names + male_names:\n",
    "    j = words_all.index(w)\n",
    "    if j >= len(common_vecs): continue\n",
    "    before = [words_all[k] for k in sims_original[j][-6:-1]]\n",
    "    after = [words_all[k] for k in sims_transformed[j][-6:-1]]\n",
    "    print(w, before, after)\n",
    "    words.append({\"Word\": w, \"Neighbors before\": \", \".join(before), \"Neighbors after\": \", \".join(after)})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 517,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7350, 3150, 4500)"
      ]
     },
     "execution_count": 517,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape[0], X_dev.shape[0], X_test.shape[0]"
   ]
  },
  {
   "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}