Hash2vec.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import re\n",
    "from sklearn.decomposition import NMF\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from sklearn.manifold import MDS\n",
    "from sklearn.cluster import SpectralClustering\n",
    "from sklearn.metrics import homogeneity_score\n",
    "from sklearn.metrics import silhouette_score\n",
    "from sklearn.metrics import completeness_score\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import normalize\n",
    " \n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from pprint import pprint\n",
    "\n",
    "from scipy.stats import pearsonr\n",
    "from statistics import mean \n",
    "from scipy.spatial.distance import pdist\n",
    "import math\n",
    "\n",
    "from collections import Counter\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "% matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 530,
   "metadata": {},
   "outputs": [],
   "source": [
    "projects = pd.read_csv('projects.csv', names=['repoId', ' ', 'repo'])\n",
    "projects.drop(columns=[' '], inplace=True)\n",
    "#projects = projects[projects['repo'].isin(sub_data['repo'])]\n",
    "project_clone = pd.read_csv('project_clones.0.csv')\n",
    "project_clone.columns = ['cloneId', 'cloneClonedFiles', 'cloneTotalFiles', 'cloneClonedPercent', \\\n",
    "                        'hostId', 'hostAffectedFiles', 'hostTotalFiles', 'hostAffectedPercent']\n",
    "\n",
    "def loadRepoNames():\n",
    "    RepoNames = {}\n",
    "    for index, field in projects.iterrows():\n",
    "        RepoNames[int(field['repoId'])] = field['repo']\n",
    "    return RepoNames\n",
    "nameDict = loadRepoNames()\n",
    "def simBetwRepos(cloneTotalFiles, ClonedPercent, hostTotalFiles, hostAffectedPercent):\n",
    "    similarity = (cloneTotalFiles * ClonedPercent * 0.01 + hostTotalFiles * hostAffectedPercent * 0.01) \\\n",
    "    / (cloneTotalFiles + hostTotalFiles)\n",
    "    return similarity\n",
    "project_clone['sim'] = project_clone.apply(lambda x: \\\n",
    "                                           simBetwRepos(x['cloneTotalFiles'], x['cloneClonedPercent'], \\\n",
    "                                                       x['hostTotalFiles'], x['hostAffectedPercent']), axis=1)\n",
    "clone = project_clone[['cloneId', 'hostId', 'sim']]\n",
    "clone.columns = ['repo1', 'repo2', 'sim']\n",
    "host = project_clone[['hostId', 'cloneId', 'sim']]\n",
    "host.columns = ['repo1', 'repo2', 'sim']\n",
    "clone_host = clone.append(host)\n",
    "\n",
    "simMatrix = clone_host.pivot_table(index='repo1', columns='repo2', values='sim').fillna(0)\n",
    "#.sort_index(inplace=True)\n",
    "#\n",
    "np.fill_diagonal(simMatrix.values, 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 490,
   "metadata": {},
   "outputs": [],
   "source": [
    "dm = (1 - simMatrix.values)\n",
    "n = dm.shape[0]\n",
    "DM = dm[np.triu_indices(n,1)]\n",
    "from scipy.cluster.hierarchy import dendrogram, linkage, cut_tree, fcluster, cophenet\n",
    "import scipy.cluster.hierarchy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grid Search for the number of clusters w.r.t silhouette score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a3016a6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cut_off = [i for i in np.arange(0.4, 1, 0.05)]\n",
    "silhouette = {}\n",
    "for co in cut_off:\n",
    "    nodes = fcluster(z, co, criterion='distance')\n",
    "    silhouette[co] = silhouette_score(dm, nodes, metric='correlation')\n",
    "plt.figure()\n",
    "plt.plot(list(silhouette.keys()), list(silhouette.values()))\n",
    "plt.xlabel(\"Cut-off Distance\")\n",
    "plt.ylabel(\"Silhouette Score\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 491,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a944d2748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20, 11))\n",
    "plt.title('Hierarchical Clustering Dendrogram')\n",
    "plt.xlabel('repo index')\n",
    "plt.ylabel('distance')\n",
    "#labelList = range(1, 9)\n",
    "z = linkage(DM)\n",
    "dendrogram(\n",
    "    z,\n",
    "   # truncate_mode='lastp',\n",
    "    leaf_font_size=10.,  # font size for the x axis labels\n",
    "    labels=simMatrix.index,\n",
    "    show_contracted=True\n",
    ")\n",
    "plt.axhline(y=.5, xmin=0, xmax=4, linewidth=2, color = 'k')\n",
    "#plt.axhline()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 487,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "105"
      ]
     },
     "execution_count": 487,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(set(fcluster(z, 0.5, criterion='distance')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 492,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_table = pd.DataFrame({'repoId': simMatrix.index, 'label': fcluster(z, 0.5, criterion='distance')})\n",
    "eval_table = eval_table.merge(projects, on='repoId')[['repoId', 'repo', 'label']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load cleaned Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 486,
   "metadata": {},
   "outputs": [],
   "source": [
    "#sub_data = pd.read_csv('stemmed_data.csv')\n",
    "tn_data = sub_data[['repo', 'stemmed_name', 'tn_length', 'tn_difflen']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 326,
   "metadata": {},
   "outputs": [],
   "source": [
    "sub_data = sub_data.fillna(' ')\n",
    "sub_data = sub_data.drop_duplicates(subset=['repo'], keep='first').reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 478,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1769.000000\n",
       "mean     1116.970605\n",
       "std      1357.940917\n",
       "min        47.000000\n",
       "25%       299.000000\n",
       "50%       642.000000\n",
       "75%      1364.000000\n",
       "max      9952.000000\n",
       "Name: tn_difflen, dtype: float64"
      ]
     },
     "execution_count": 478,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tn_data['tn_difflen'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 430,
   "metadata": {},
   "outputs": [],
   "source": [
    "sample = pd.read_csv('sample.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 432,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['repo', 'token.name', 'token.comment', 'tag', 'stemmed_name',\n",
       "       'stemmed_comment', 'tn_length', 'tc_length', 'tn_difflen',\n",
       "       'tc_difflen'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 432,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 439,
   "metadata": {},
   "outputs": [],
   "source": [
    "sub_data = sub_data.append(sample[['repo', 'token.name', 'token.comment', 'tn_length', 'tc_length',\n",
    "       'tn_difflen', 'tc_difflen', 'stemmed_name', 'stemmed_comment']]).drop_duplicates(subset=['repo'], keep='first').reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 443,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1785, 4)"
      ]
     },
     "execution_count": 443,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tn_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## hash embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def hash2vec(doc, length):\n",
    "    doc = doc.split(' ')\n",
    "    doc = Counter(doc)\n",
    "    embedding = [0] * length\n",
    "    for word in doc:\n",
    "        hval = hash(word) % length\n",
    "        embedding[hval] += doc[word]\n",
    "    return np.array(embedding)\n",
    "\n",
    "def vectorize(column):\n",
    "    '''Transfer a column of dataframe into np.array'''\n",
    "    fit_vec = np.array(column[0])\n",
    "    for field in column:\n",
    "        fit_vec = np.vstack((fit_vec, field))\n",
    "    fit_vec = np.delete(fit_vec, 0, 0)\n",
    "    return fit_vec\n",
    "\n",
    "def pearson_dist(matrix):\n",
    "    '''Calculate distance matrix using pearson correlation'''\n",
    "    n, m = matrix.shape\n",
    "    value = pdist(matrix, 'correlation')\n",
    "    start = 0\n",
    "    pearson_affinity = np.zeros((n, n))\n",
    "    for i in range(n - 1):\n",
    "        for j in range(i + 1, n):\n",
    "            pearson_affinity[i, j] = value[start + j - 1]\n",
    "            pearson_affinity[j, i] = value[start + j - 1]\n",
    "        start += n - i - 2\n",
    "    return pearson_affinity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [],
   "source": [
    "def is_prime(num):\n",
    "    \"\"\"Returns True if the number is prime\n",
    "    else False.\"\"\"\n",
    "    if num == 0 or num == 1:\n",
    "        return False\n",
    "    for x in range(2, num):\n",
    "        if num % x == 0:\n",
    "            return False\n",
    "    else:\n",
    "        return True\n",
    "buckets_length = []\n",
    "for i in filter(is_prime, range(500, 1500, 23)):\n",
    "    buckets_length.append(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 523\n",
      "Finished calculating k = 140 for bucket length = 523\n",
      "Finished calculating k = 180 for bucket length = 523\n",
      "Finished calculating k = 220 for bucket length = 523\n",
      "Finished calculating k = 260 for bucket length = 523\n",
      "Finished calculating k = 300 for bucket length = 523\n",
      "Finished calculating k = 340 for bucket length = 523\n",
      "Finished calculating k = 380 for bucket length = 523\n",
      "Finished calculating k = 420 for bucket length = 523\n",
      "Finished calculating k = 460 for bucket length = 523\n",
      "Finished calculating k = 500 for bucket length = 523\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 569\n",
      "Finished calculating k = 140 for bucket length = 569\n",
      "Finished calculating k = 180 for bucket length = 569\n",
      "Finished calculating k = 220 for bucket length = 569\n",
      "Finished calculating k = 260 for bucket length = 569\n",
      "Finished calculating k = 300 for bucket length = 569\n",
      "Finished calculating k = 340 for bucket length = 569\n",
      "Finished calculating k = 380 for bucket length = 569\n",
      "Finished calculating k = 420 for bucket length = 569\n",
      "Finished calculating k = 460 for bucket length = 569\n",
      "Finished calculating k = 500 for bucket length = 569\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 661\n",
      "Finished calculating k = 140 for bucket length = 661\n",
      "Finished calculating k = 180 for bucket length = 661\n",
      "Finished calculating k = 220 for bucket length = 661\n",
      "Finished calculating k = 260 for bucket length = 661\n",
      "Finished calculating k = 300 for bucket length = 661\n",
      "Finished calculating k = 340 for bucket length = 661\n",
      "Finished calculating k = 380 for bucket length = 661\n",
      "Finished calculating k = 420 for bucket length = 661\n",
      "Finished calculating k = 460 for bucket length = 661\n",
      "Finished calculating k = 500 for bucket length = 661\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 937\n",
      "Finished calculating k = 140 for bucket length = 937\n",
      "Finished calculating k = 180 for bucket length = 937\n",
      "Finished calculating k = 220 for bucket length = 937\n",
      "Finished calculating k = 260 for bucket length = 937\n",
      "Finished calculating k = 300 for bucket length = 937\n",
      "Finished calculating k = 340 for bucket length = 937\n",
      "Finished calculating k = 380 for bucket length = 937\n",
      "Finished calculating k = 420 for bucket length = 937\n",
      "Finished calculating k = 460 for bucket length = 937\n",
      "Finished calculating k = 500 for bucket length = 937\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 983\n",
      "Finished calculating k = 140 for bucket length = 983\n",
      "Finished calculating k = 180 for bucket length = 983\n",
      "Finished calculating k = 220 for bucket length = 983\n",
      "Finished calculating k = 260 for bucket length = 983\n",
      "Finished calculating k = 300 for bucket length = 983\n",
      "Finished calculating k = 340 for bucket length = 983\n",
      "Finished calculating k = 380 for bucket length = 983\n",
      "Finished calculating k = 420 for bucket length = 983\n",
      "Finished calculating k = 460 for bucket length = 983\n",
      "Finished calculating k = 500 for bucket length = 983\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1213\n",
      "Finished calculating k = 140 for bucket length = 1213\n",
      "Finished calculating k = 180 for bucket length = 1213\n",
      "Finished calculating k = 220 for bucket length = 1213\n",
      "Finished calculating k = 260 for bucket length = 1213\n",
      "Finished calculating k = 300 for bucket length = 1213\n",
      "Finished calculating k = 340 for bucket length = 1213\n",
      "Finished calculating k = 380 for bucket length = 1213\n",
      "Finished calculating k = 420 for bucket length = 1213\n",
      "Finished calculating k = 460 for bucket length = 1213\n",
      "Finished calculating k = 500 for bucket length = 1213\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1259\n",
      "Finished calculating k = 140 for bucket length = 1259\n",
      "Finished calculating k = 180 for bucket length = 1259\n",
      "Finished calculating k = 220 for bucket length = 1259\n",
      "Finished calculating k = 260 for bucket length = 1259\n",
      "Finished calculating k = 300 for bucket length = 1259\n",
      "Finished calculating k = 340 for bucket length = 1259\n",
      "Finished calculating k = 380 for bucket length = 1259\n",
      "Finished calculating k = 420 for bucket length = 1259\n",
      "Finished calculating k = 460 for bucket length = 1259\n",
      "Finished calculating k = 500 for bucket length = 1259\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1489\n",
      "Finished calculating k = 140 for bucket length = 1489\n",
      "Finished calculating k = 180 for bucket length = 1489\n",
      "Finished calculating k = 220 for bucket length = 1489\n",
      "Finished calculating k = 260 for bucket length = 1489\n",
      "Finished calculating k = 300 for bucket length = 1489\n",
      "Finished calculating k = 340 for bucket length = 1489\n",
      "Finished calculating k = 380 for bucket length = 1489\n",
      "Finished calculating k = 420 for bucket length = 1489\n",
      "Finished calculating k = 460 for bucket length = 1489\n",
      "Finished calculating k = 500 for bucket length = 1489\n"
     ]
    }
   ],
   "source": [
    "sse = {}\n",
    "for b_len in buckets_length:\n",
    "    tn_data['hash2vec'] = tn_data['stemmed_name'].apply(lambda x: hash2vec(x, b_len))\n",
    "    print(\"Hashing finished...\")\n",
    "    name_fit = vectorize(tn_data['hash2vec'])\n",
    "    print('vectorization finished...')\n",
    "    sse_b = {}\n",
    "    key = 'bucket' + str(b_len)\n",
    "    for k in range(100, 520, 40):\n",
    "        kmeans = KMeans(n_clusters=k, max_iter=200, random_state=122).fit(name_fit)\n",
    "        #data[\"clusters\"] = kmeans.labels_\n",
    "        #print(data[\"clusters\"])\n",
    "        sse_b[k] = kmeans.inertia_ # Inertia: Sum of distances of samples to their closest cluster center\n",
    "        print(\"Finished calculating k = \" + str(k) + \" for bucket length = \" + str(b_len))\n",
    "    sse[key] = sse_b\n",
    "#plt.figure()\n",
    "#plt.plot(list(sse.keys()), list(sse.values()))\n",
    "#plt.xlabel(\"Number of cluster\")\n",
    "#plt.ylabel(\"SSE\")\n",
    "#plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2f661eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "n_cluster = [i for i in range(100, 520, 40)]\n",
    "for n_bucket, record in sse.items():\n",
    "    plt.plot(n_cluster, list(record.values()), label=str(n_bucket))\n",
    "plt.xlabel(\"Number of Cluster\")\n",
    "plt.ylabel(\"SSE\")\n",
    "plt.legend(title='Length of Hash Table', loc='best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 506,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "tn_data['hash2vec'] = tn_data['stemmed_name'].apply(lambda x: hash2vec(x, 523))\n",
    "name_fit = vectorize(tn_data['hash2vec'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 494,
   "metadata": {},
   "outputs": [],
   "source": [
    "testset = tn_data[tn_data['repo'].isin(eval_table['repo'])]\n",
    "test_fit = vectorize(testset['hash2vec'])\n",
    "trainset = tn_data[~tn_data['repo'].isin(eval_table['repo'])].reset_index()\n",
    "train_fit = vectorize(trainset['hash2vec'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## AgglomerativeClustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 495,
   "metadata": {},
   "outputs": [],
   "source": [
    "#cluster_ac = AgglomerativeClustering(n_clusters=200, linkage='average',\n",
    "#                           affinity=pearson_dist)\n",
    "#cluster_ac.fit(name_fit)\n",
    "Z = linkage(name_fit, metric = 'correlation')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2fe75a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cut_off = [i for i in np.arange(0.1, 0.7, 0.1)]\n",
    "silhouette = {}\n",
    "dm = pearson_dist(name_fit)\n",
    "for co in cut_off:\n",
    "    nodes = fcluster(Z, co, criterion='distance')\n",
    "    silhouette[co] = silhouette_score(dm, nodes, metric='correlation')\n",
    "plt.figure()\n",
    "plt.plot(list(silhouette.keys()), list(silhouette.values()))\n",
    "plt.xlabel(\"Cut-off Distance\")\n",
    "plt.ylabel(\"Silhouette Score\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.04704076463793034"
      ]
     },
     "execution_count": 335,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nodes = fcluster(Z, 0.32, criterion='distance')\n",
    "silhouette_score(dm, nodes, metric='euclidean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 484,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2c653e80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "scipy.cluster.hierarchy.set_link_color_palette(['m', 'c', 'y', 'k'])\n",
    "plt.figure(figsize=(20, 11))\n",
    "plt.title('Hierarchical Clustering Dendrogram')\n",
    "plt.xlabel('repo index')\n",
    "plt.ylabel('distance')\n",
    "\n",
    "dendrogram(\n",
    "    Z,\n",
    "    #truncate_mode='lastp',\n",
    "   # leaf_font_size=10.,  # font size for the x axis labels\n",
    "   # labels=,\n",
    "    show_contracted=True,\n",
    "    orientation='top'\n",
    "   # above_threshold_color='y'\n",
    ")\n",
    "plt.axhline(y=.5, xmin=0, xmax=4, linewidth=4, color = '#d95f02')\n",
    "#plt.axhline()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 499,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "hac_data = tn_data\n",
    "hac_data['pred_label'] = fcluster(Z, 0.6, criterion='distance')\n",
    "test_hac = hac_data[hac_data['repo'].isin(eval_table['repo'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 497,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1372"
      ]
     },
     "execution_count": 497,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(set(hac_data['pred_label']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 500,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.061\n",
      "Completeness: 0.653\n",
      "Recall is 0.93\n",
      "Precision is 0.33\n"
     ]
    }
   ],
   "source": [
    "eval_ac = eval_table[['repo', 'label']].merge(test_hac[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_ac['label'], eval_ac['pred_label'])\n",
    "completeness = completeness_score(eval_ac['label'], eval_ac['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_ac.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_ac.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Explained variance ratio:  [0.33594388 0.09080327]\n"
     ]
    }
   ],
   "source": [
    "pca = PCA(n_components=2)\n",
    "pca_res = pca.fit_transform(name_fit)\n",
    "x = pca_res[:, 0]\n",
    "y = pca_res[:, 1]\n",
    "print(\"Explained variance ratio: \", pca.explained_variance_ratio_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1a2e5d8ba8>"
      ]
     },
     "execution_count": 345,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2e417e48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot\n",
    "plt.figure(figsize=(8, 8))\n",
    "plt.scatter(x, y, c=hac_data['pred_label'])\n",
    "plt.ylabel('Component 2')\n",
    "plt.xlabel('Component 1')\n",
    "plt.title(\"Segregation of Topic Clusters\", )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## KMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 572,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished calculating k = 50\n",
      "Finished calculating k = 100\n",
      "Finished calculating k = 150\n",
      "Finished calculating k = 200\n",
      "Finished calculating k = 250\n",
      "Finished calculating k = 300\n",
      "Finished calculating k = 350\n",
      "Finished calculating k = 400\n",
      "Finished calculating k = 450\n",
      "Finished calculating k = 500\n",
      "Finished calculating k = 550\n",
      "Finished calculating k = 600\n",
      "Finished calculating k = 650\n",
      "Finished calculating k = 700\n",
      "Finished calculating k = 750\n",
      "Finished calculating k = 800\n",
      "Finished calculating k = 850\n",
      "Finished calculating k = 900\n",
      "Finished calculating k = 950\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2d5a5278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sse = {}\n",
    "for k in range(50, 1000, 50):\n",
    "    kmeans = KMeans(n_clusters=k, max_iter=1000).fit(name_fit)\n",
    "    #data[\"clusters\"] = kmeans.labels_\n",
    "    #print(data[\"clusters\"])\n",
    "    sse[k] = kmeans.inertia_ # Inertia: Sum of distances of samples to their closest cluster center\n",
    "    print(\"Finished calculating k = \" + str(k))\n",
    "plt.figure()\n",
    "plt.plot(list(sse.keys()), list(sse.values()))\n",
    "plt.xlabel(\"Number of cluster\")\n",
    "plt.ylabel(\"SSE\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "100"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans.fit_predict(test_fit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 510,
   "metadata": {},
   "outputs": [],
   "source": [
    "kmeans_pearson = 1 - pearson_dist(name_fit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 512,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "kmeans_data = tn_data\n",
    "kmeans = KMeans(n_clusters=150, random_state=122, precompute_distances=True).fit(kmeans_pearson)\n",
    "kmeans_data['pred_label'] = kmeans.labels_\n",
    "kmeans_test = kmeans_data[kmeans_data['repo'].isin(eval_table['repo'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 545,
   "metadata": {},
   "outputs": [],
   "source": [
    "kmeans_eval = eval_table[eval_table['repo'].isin(kmeans_test['repo'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 546,
   "metadata": {},
   "outputs": [],
   "source": [
    "kmeans_eval = kmeans_eval.merge(tn_data[['repo', 'stemmed_name']], on='repo').drop_duplicates(subset=['repo'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 547,
   "metadata": {},
   "outputs": [],
   "source": [
    "kmeans_eval['hash2vec'] = kmeans_eval['stemmed_name'].apply(lambda x: hash2vec(x, 523))\n",
    "test_fit = vectorize(kmeans_eval['hash2vec'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 513,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.864\n",
      "Completeness: 0.731\n",
      "Recall is 0.07\n",
      "Precision is 0.55\n"
     ]
    }
   ],
   "source": [
    "#test_kmeans = testset\n",
    "#kmeans_test = KMeans(n_clusters=77, max_iter=1000).fit(test_fit)\n",
    "#test_kmeans['pred_label'] = kmeans_test.labels_\n",
    "\n",
    "eval_kmeans = eval_table[['repo', 'label']].merge(kmeans_test[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_kmeans['label'], eval_kmeans['pred_label'])\n",
    "completeness = completeness_score(eval_kmeans['label'], eval_kmeans['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_kmeans.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_kmeans.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.739\n",
      "Completeness: 0.543\n",
      "Recall is 0.42\n",
      "Precision is 0.59\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "test_kmeans = testset\n",
    "kmeans_test = KMeans(n_clusters=77, max_iter=1000).fit(test_fit)\n",
    "test_kmeans['pred_label'] = kmeans_test.labels_\n",
    "\n",
    "eval_kmeans = eval_table[['repo', 'label']].merge(test_kmeans[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_kmeans['label'], eval_kmeans['pred_label'])\n",
    "completeness = completeness_score(eval_kmeans['label'], eval_kmeans['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_kmeans.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_kmeans.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1a2fa8b5c0>"
      ]
     },
     "execution_count": 350,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2fee09e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8, 8))\n",
    "plt.scatter(x, y, c=kmeans_data['pred_label'])\n",
    "plt.ylabel('Component 2')\n",
    "plt.xlabel('Component 1')\n",
    "plt.title(\"Segregation of Topic Clusters\", )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 557,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('label', 'r') as f:\n",
    "    lda_label = [i.split('\\n')[0] for i in f.readlines()] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 564,
   "metadata": {},
   "outputs": [],
   "source": [
    "lda_label = pd.read_csv('label', header=None, names=['lda_label'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 569,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6098520692057462"
      ]
     },
     "execution_count": 569,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "homogeneity_score(kmeans_data['pred_label'], lda_label['lda_label'] )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 571,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5747475927160394"
      ]
     },
     "execution_count": 571,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "completeness_score(kmeans_data['pred_label'], lda_label['lda_label'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Normalize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 514,
   "metadata": {},
   "outputs": [],
   "source": [
    "name_fit_norm = normalize(name_fit, axis=1, norm='l2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 352,
   "metadata": {},
   "outputs": [],
   "source": [
    "buckets_length = []\n",
    "for i in filter(is_prime, range(500, 1500, 23)):\n",
    "    buckets_length.append(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 353,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[523, 569, 661, 937, 983, 1213, 1259, 1489]"
      ]
     },
     "execution_count": 353,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "buckets_length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 523\n",
      "Finished calculating k = 140 for bucket length = 523\n",
      "Finished calculating k = 180 for bucket length = 523\n",
      "Finished calculating k = 220 for bucket length = 523\n",
      "Finished calculating k = 260 for bucket length = 523\n",
      "Finished calculating k = 300 for bucket length = 523\n",
      "Finished calculating k = 340 for bucket length = 523\n",
      "Finished calculating k = 380 for bucket length = 523\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 569\n",
      "Finished calculating k = 140 for bucket length = 569\n",
      "Finished calculating k = 180 for bucket length = 569\n",
      "Finished calculating k = 220 for bucket length = 569\n",
      "Finished calculating k = 260 for bucket length = 569\n",
      "Finished calculating k = 300 for bucket length = 569\n",
      "Finished calculating k = 340 for bucket length = 569\n",
      "Finished calculating k = 380 for bucket length = 569\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 661\n",
      "Finished calculating k = 140 for bucket length = 661\n",
      "Finished calculating k = 180 for bucket length = 661\n",
      "Finished calculating k = 220 for bucket length = 661\n",
      "Finished calculating k = 260 for bucket length = 661\n",
      "Finished calculating k = 300 for bucket length = 661\n",
      "Finished calculating k = 340 for bucket length = 661\n",
      "Finished calculating k = 380 for bucket length = 661\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 937\n",
      "Finished calculating k = 140 for bucket length = 937\n",
      "Finished calculating k = 180 for bucket length = 937\n",
      "Finished calculating k = 220 for bucket length = 937\n",
      "Finished calculating k = 260 for bucket length = 937\n",
      "Finished calculating k = 300 for bucket length = 937\n",
      "Finished calculating k = 340 for bucket length = 937\n",
      "Finished calculating k = 380 for bucket length = 937\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 983\n",
      "Finished calculating k = 140 for bucket length = 983\n",
      "Finished calculating k = 180 for bucket length = 983\n",
      "Finished calculating k = 220 for bucket length = 983\n",
      "Finished calculating k = 260 for bucket length = 983\n",
      "Finished calculating k = 300 for bucket length = 983\n",
      "Finished calculating k = 340 for bucket length = 983\n",
      "Finished calculating k = 380 for bucket length = 983\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1213\n",
      "Finished calculating k = 140 for bucket length = 1213\n",
      "Finished calculating k = 180 for bucket length = 1213\n",
      "Finished calculating k = 220 for bucket length = 1213\n",
      "Finished calculating k = 260 for bucket length = 1213\n",
      "Finished calculating k = 300 for bucket length = 1213\n",
      "Finished calculating k = 340 for bucket length = 1213\n",
      "Finished calculating k = 380 for bucket length = 1213\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1259\n",
      "Finished calculating k = 140 for bucket length = 1259\n",
      "Finished calculating k = 180 for bucket length = 1259\n",
      "Finished calculating k = 220 for bucket length = 1259\n",
      "Finished calculating k = 260 for bucket length = 1259\n",
      "Finished calculating k = 300 for bucket length = 1259\n",
      "Finished calculating k = 340 for bucket length = 1259\n",
      "Finished calculating k = 380 for bucket length = 1259\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1489\n",
      "Finished calculating k = 140 for bucket length = 1489\n",
      "Finished calculating k = 180 for bucket length = 1489\n",
      "Finished calculating k = 220 for bucket length = 1489\n",
      "Finished calculating k = 260 for bucket length = 1489\n",
      "Finished calculating k = 300 for bucket length = 1489\n",
      "Finished calculating k = 340 for bucket length = 1489\n",
      "Finished calculating k = 380 for bucket length = 1489\n"
     ]
    }
   ],
   "source": [
    "sse = {}\n",
    "for b_len in buckets_length:\n",
    "    tn_data['hash2vec'] = tn_data['stemmed_name'].apply(lambda x: hash2vec(x, b_len))\n",
    "    print(\"Hashing finished...\")\n",
    "    name_fit = vectorize(tn_data['hash2vec'])\n",
    "    name_fit_norm = normalize(name_fit, axis=1, norm='l2')\n",
    "    print('vectorization finished...')\n",
    "    sse_b = {}\n",
    "    key = 'bucket' + str(b_len)\n",
    "    for k in range(100, 400, 40):\n",
    "        kmeans = KMeans(n_clusters=k, random_state=122, max_iter=200).fit(name_fit_norm)\n",
    "        #data[\"clusters\"] = kmeans.labels_\n",
    "        #print(data[\"clusters\"])\n",
    "        sse_b[k] = kmeans.inertia_ # Inertia: Sum of distances of samples to their closest cluster center\n",
    "        print(\"Finished calculating k = \" + str(k) + \" for bucket length = \" + str(b_len))\n",
    "    sse[key] = sse_b\n",
    "#plt.figure()\n",
    "#plt.plot(list(sse.keys()), list(sse.values()))\n",
    "#plt.xlabel(\"Number of cluster\")\n",
    "#plt.ylabel(\"SSE\")\n",
    "#plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2a110780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "n_cluster = [i for i in range(100, 400, 40)]\n",
    "for n_bucket, record in sse.items():\n",
    "    plt.plot(n_cluster, list(record.values()), label=str(n_bucket))\n",
    "plt.xlabel(\"Number of Cluster\")\n",
    "plt.ylabel(\"SSE\")\n",
    "plt.legend(title='Length of Hash Table', loc='best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hierachy clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 354,
   "metadata": {},
   "outputs": [],
   "source": [
    "Z_norm = linkage(name_fit_norm, metric = 'correlation')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 355,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJIAAAKWCAYAAAASrLMzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAIABJREFUeJzs3X+YZAld3/vPd05T0kU11cbFsQVkcVhGF6KF7B0lt810EqOg/AomBiYmD0js6R5HnasyFwzhAgkk7MOakIRcHX284o0tECK6PMKDuegU2yLCEIsfC7IMBB2g08BCn+3aLfbUnPO9f5w61dU93dOnf1Sf+vF+PU891V11qurb1bWzs5/9fr/H3F0AAAAAAADAbo4VXQAAAAAAAACGA0ESAAAAAAAAciFIAgAAAAAAQC4ESQAAAAAAAMiFIAkAAAAAAAC5ECQBAAAAAAAgF4IkAABGjJnda2ZzA1DHrWbmZjaxw/2/bGa/0c/XyPH4V5vZfzlIDYfFzJpm9p1F13EYOr+TJxVdBwAAOHwESQAADBEz+7yZ/dCW215sZsvZ9+7+FHe/fOTF7ZG7v97d/3m/X8fMzpjZlU5Qs2Jm7zGz2UN8/gOFWRl3r7j75w6rrkwnLGub2Xrncp+Z/Sczmzns1wIAAKOPIAkAAEiS9hOEmFnQj1oOi5n9gqR/L+n1ko5L+g5J/1nS84qsq9dBA6ic3ubuU5L+hqR/IOnbJH2kiDDpMD8zluLvswAAHCH+xQsAwIjp7Voys2Nm9nIz+6yZ3W9mbzezv9G5L+ukeamZ/bWkP+7c/l/N7H+ZWWhm7zezp/Q892+Z2f9tZu82swcl/R0zmzSzu8zsrzqPWTazyZ6S/omZ/bWZfdXM/kXPc20aKzOzWTP7gJmtmdk1M3tx5/YfM7O/MLMHOre/Ouf7UJX0Wkk/4+6/5+4Punvb3d/l7i/b5vg5M/vCTd7LU53OpgfMbNXMfqVz2Ps712udrqdndI7/KTP7lJl93czea2ZP6HleN7OfMbPPSPpMz21P6nmf32xmf9jpIvpzMzvR8/gfNrNPd97v/2xmdTPbtbur8/PfK+kfS/qKpF/sec5nm1mj8/5/wMy+Z8v78Etm9rHOa77NzB7Zc//LOt1eXzKzn9ryHm73mama2W+b2Vc6n5tXZoGQmQWdz9NXzex/mtl56+n4MrPLZvY6M/tTSQ9J+k4ze0nnvV43s8+Z2dmtv1czu2hmX+7U+Xwz+1FLu7O+Zma/vNt7BwAAUgRJAACMtp+V9HxJpyV9u6SvS3rzlmNOS/puST/S+f49km6T9K2S/oek39ly/BlJr5M0JWlZ0hslPV3S31La8XJRUtJz/Kykk5L+nqRXmdl3by2yE7K8R9J/lPQYSTVJjc7dD0r6Z5KmJf2YpEUze36On/0Zkh4p6Z05js3jTZLe5O6PlnRC0ts7t//tzvV0Zzztz8zseZJ+WdILlP4890j63S3P93xJ3y/p9h1e74WSXiPpmyVdVfqey8xukfQOSa+Q9C2SPq30vc/N3WNJfyDpBzvP+TRJvynpbOc5f03S3Wb2TT0P+wlJz5T0REnfI+nFncc+U9IvSfr7Sj83m0YvO7Z+Zv6jpKqk71T6+ftnkl7SOfanJT1L6Wfg+5S+T1v9U0nznef7K0lflvRsSY/uPM+/M7Pv6zn+25R+Fh4r6VWSfl3STyr93P6gpH9pZk/c6f0CAAAbCJIAABg+v9/pGlkzszWlo1o7WZD0L9z9C+7+sKRXS/qHtnmc6tWdbp2WJLn7b7r7es/x39vp7sn8gbv/qbsnkiJJPyXp5939i+4eu/sHOo/NvMbdW+7+UUkflfS929R5RtL/5+6/2+maud/dG516Lrv7x909cfePKQ1kTud4n75F0lfd/XqOY/NoS3qSmd3i7k13/+BNjl2Q9G/c/VOd13+9pFpvV1Ln/q9l7/s23unuH+o8/neUBiuS9KOS7u10WV2X9B8k/a99/DxfUhr8SWko82vu/ued3+FbJD0s6Qd6jv8P7v4ld/+apHf11PMTkv4fd/+Euz+o9DOzVe9npq00JHtF53P2eUl3KQ2Hsud7U+cz+3VJ/3ab5/std7/X3a93Pi9/6O6f9VRd0h+pE5J1tCW9zt3bkt4q6ZbOa6x3OrQ+qe0/lwAAYAuCJAAAhs/z3X06u0g6d5NjnyDpnT2h06ckxUr3BWWuZV90xor+raWjcA9I+nznrlu2O75z+yMlffYmNfSGHA9JqmxzzON3eg4z+34z+5POGFSoNKS5Zbtjt7hf0i12eDuIXirpyZL+0sw+bGbPvsmxT5D0pp73/WuSTGlHTObato/csNP79u29j3V3l7RpJC+nx3bqyur9xS0B5eM7r7WnepR2CG219TPziC3H/ZU23putz7fd+7TpNjN7lpl9sDOmtqY0bOv9jNzf6cKSpCy4W+25v6XtP5cAAGALgiQAAEbbNUnP6g2e3P2R7v7FnmO85+szShdR/5DS0aNbO7fbDsd/VdI3lI56HbTOnZ5jSdLdkh7v7lVJv7qlnp38mdKumjxjcFI6QlfOvrF0KfRjsu/d/TPu/iKlI39vkPQOM3uUNr8fmWuSzm553yfd/QM9x2z3uDxWJD2up07r/T6Pzj6i5ygducvqfd2WesvuvnUcb6d6Ht/z/Xdsc8zWz0xbaXjV+5jsM7np59vy3Dc8X2f87r8pHbE83glX3618nxEAALBHBEkAAIy2X5X0umykyswe09nfs5MppeHL/UpDldff7Mk7o0q/KelXzOzbOx1Nz9iyWyeP35H0Q2b2E2Y2YWbfYmbZ6NSUpK+5+zfM7JTSsGtX7h4q3Yfz5s5y5bKZPaLTvXLnNg+5T9IjLV3u/QhJr5TU/TnM7CfN7DGdn3mtc3OidGl1onTfT+ZXJb3COovKO8ul/1HO92I3fyjpb3Z+pglJP6N0B9CuOu/tdysdD/w2SdnC8F+XtNDp/jIze1TnfZjK8bRvl/RiM7vdzMqS/q+bHdzpDHq70s/lVOez+QuSssXrb5f082b2WDOblvR/7vL6JaW/p69Ium5mz5L0wznqBgAA+0CQBADAaHuT0m6ePzKzdUkfVLrgeSe/rXTM6ItK98bcbA9Q5pckfVzSh5WOSr1Be/w7hrv/tdJxpF/sPEdDGztrzkl6baf+V2ljyXWe571LaUjxSqVBwzVJ5yX9/jbHhp3X+g2lP/+D2jwy9kxJ95pZU+n7+sLO7qeHlC6S/tPOWNgPuPs7lb4Pb+2MCH5C6QLpA3P3r0r6R5LuVBr43S7pitIAcCf/uFN3qPTzcL+kp7v7lzrPeUXpkuv/pHQh+1V1lmnnqOc9kv690rP+Xe1c7+Znlb6/n1O6fHtJaSAppaHWH0n6mKS/UNpddF3pSOZ2r78u6eeUfi6+rjRovDtP7QAAYO8sHasHAADAMOqMqX1B0j9x9z8pup7D1ukw+lV3f8KuBwMAgL6jIwkAAGDImNmPmNl0Z4Twl5XuA8rTPTbwzGzSzH60M4b3WKWjcu8sui4AAJAiSAIAABg+z1B6lruvKl2a/Xx3b938IUPDJL1G6ZjaXyg90+CrCq0IAAB0MdoGAAAAAACAXOhIAgAAAAAAQC4ESQAAAAAAAMhlougC9uqWW27xW2+9tegyAAAAAAAARsZHPvKRr7r7Y3Y7buiCpFtvvVVXrlwpugwAAAAAAICRYWZ/lec4RtsAAAAAAACQC0ESAAAAAAAAciFIAgAAAAAAQC4ESQAAAAAAAMiFIAkAAAAAAAC5ECQBAAAAAAAgF4IkAAAAAAAA5EKQBAAAAAAAgFwIkgAAAAAAAJALQRIAAAAAAAByIUgCAAAAAABALgRJAAAAAAAAyIUgCQAAAAAAALkQJAEAAAAAACAXgiQAAAAAAADkQpAEAAAAAACAXAiSAAAAAAAAkAtBEgAAAAAAAHIhSAIAAAAAAEAuBEkAAAAAAADIhSAJAAAAAAAAuRAkAQAAAAAAIBeCJAAAAAAAAORCkAQAAAAAAIBcCJIAAAAAAACQy9AFSZ/+tDQ3J126VHQlAAAAAAAA42XogqRWS2o0pKWloisBAAAAAAAYL0MXJE1OSrVa0VUAAAAAAACMn6ELkgAAAAAAAFAMgiQAAAAAAADkMlF0AXvVbqc7ko4fL7oSAAAAAACA8TJ0HUnXr0thKM3MFF0JAAAAAADAeBm6IAkAAAAAAADFGNogqdGQLl0qugoAAAAAAIDxMZRBUrWajrctLRVdCQAAAAAAwPgYyiAJAAAAAAAAR48gCQAAAAAAALkQJAEAAAAAACAXgiQAAAAAAADkQpAEAAAAAACAXIYySIqi9LrRkC5dKrYWAAAAAACAcTG0QVIQpF8vLRVbCwAAAAAAwLgYyiBJkioVqVYrugoAAAAAAIDxMbRBUhRJH/oQ420AAAAAAABHZaLoAvYriqQ4llot6eLF9Lb5+WJrAgAAAAAAGGVD25EkpXuS2JUEAAAAAABwNIY6SJLYlQQAAAAAAHBUhj5IAgAAAAAAwNEgSAIAAAAAAEAuBEkAAAAAAADIhSAJAAAAAAAAuRAkAQAAAAAAIBeCJAAAAAAAAORCkAQAAAAAAIBcCJIAAAAAAACQC0ESAAAAAAAAciFIAgAAAAAAQC59DZLM7Jlm9mkzu2pmL9/m/n9nZo3O5T4zW+tnPQAAAAAAANi/iX49sZkFkt4s6e9L+oKkD5vZ3e7+yewYd/8/eo7/WUlP61c9AAAAAAAAOJh+diSdknTV3T/n7pGkt0p63k2Of5Gk3+1jPQAAAAAAADiAfgZJj5V0ref7L3Ruu4GZPUHSEyX98V5fJIqkRiO9XLq0rzoBAAAAAACQw6As236hpHe4e7zdnWY2b2ZXzOxKkvim+6Jo4+ulpX6WCAAAAAAAMN76GSR9UdLje75/XOe27bxQNxlrc/dL7n6Hu99x7JjdcP/x4+n1ysp+SwUAAAAAAMBu+hkkfVjSbWb2RDMrKQ2L7t56kJl9l6RvlvRn+32h1VUpDNNrxtsAAAAAAAD6o29Bkrtfl3Re0nslfUrS2939XjN7rZk9t+fQF0p6q7v7ds+T1+RkGiZdvEiYBAAAAAAA0A92wPzmyE1M3OHSlU23VSpSs7nxda0mXb589LUBAAAAAAAMIzP7iLvfsdtxg7Js+1CUSuk1u5IAAAAAAAAO30gFSVGUjrfNzBRdCQAAAAAAwOgZqSAJAAAAAAAA/TOSQVKjwcJtAAAAAACAwzZyQRJnbwMAAAAAAOiPkQuSokgKgvTrpaViawEAAAAAABglIxckSVKlItVqRVcBAAAAAAAwWkYySAIAAAAAAMDhI0gCAAAAAABALgRJAAAAAAAAyIUgCQAAAAAAALkQJAEAAAAAACCXkQySokhqNNLLpUtFVwMAAAAAADAaJoouoB+iKL1utaSLF9Ov5+eLqwcAAAAAAGAUjGRHkiSVSul1GEqLi9LcHN1JAAAAAAAABzGSHUmZIJDiWEoSqV5PR90kupMAAAAAAAD2Y6SDJCkNkyTJPe1OYtQNAAAAAABgf0Z2tG0rM2lyMg2Tzp1j1A0AAAAAAGCvxiZIktK9SVmHUqMhLS0VWw8AAAAAAMAwGasgSZIqFenEifTMbo0GXUkAAAAAAAB5jfyOpO2srkqtVnphZxIAAAAAAEA+Y9eRlAkC6dixdGfS4iI7kwAAAAAAAHYzlh1JGbM0UIpjqV6XlpfT2+lOAgAAAAAAuNHYdiRlKpWNBdxxnI660ZkEAAAAAABwo7EPkjJBIFWr6agbZ3MDAAAAAAC4EUFSjyhKrzmbGwAAAAAAwI0IknpE0cYC7rvuKroaAAAAAACAwUKQtIVZej0zU2wdAAAAAAAAg4YgaQeMtwEAAAAAAGxGkLQNlm4DAAAAAADciCBpG9nS7ZWVYusAAAAAAAAYJARJ28iCJPYkAQAAAAAAbCBIugn2JAEAAAAAAGwgSNpBELAnCQAAAAAAoBdB0g7c0+t77pHm5uhMAgAAAAAAmCi6gEGVJBvX9Xo65iZJ8/PF1QQAAAAAAFAkOpJuIgg2LhJjbgAAAAAAYLwRJOVQKqVncmP5NgAAAAAAGGcESTlEkdRqsXwbAAAAAACMN4KknKrV9JquJAAAAAAAMK4IkvagWqUrCQAAAAAAjC+CpD2IovSariQAAAAAADCOCJL2IIo4gxsAAAAAABhfBEl7VCql1ysrxdYBAAAAAABw1AiS9iiK0j1JV69K5bI0N8eYGwAAAAAAGA8ESfsQBFKSSK2WVK8z5gYAAAAAAMYDQdI+BUF6FjcAAAAAAIBxQZB0AJzFDQAAAAAAjJOhDpLcpTjeCHSOGmdxAwAAAAAA42SogySz9LqoIEmSKhWpVivu9QEAAAAAAI7KRNEFDLsoSkfbjh8vuhIAAAAAAID+GuqOpEEQRVIYSjMzRVcCAAAAAADQXwRJh2R5WZqbY+k2AAAAAAAYXQRJhyAI0qXf9bq0uChNTxMqAQAAAACA0UOQdEiCIL0kibS+noZKCwtSuZwGS4RKAAAAAABg2BEkHaJKRapWpampNFRyl1qtdIfS0lLR1QEAAAAAABwMQVJOUZSOr7nnO75S2ehSkqR77pFKJTqUAAAAAADA8CJIyimK0muzvT82G3lrtzc6lBYX2aMEAAAAAACGy0TRBYyLrDMpjtPrJEn3KDUa6ffz88XUBQAAAAAAkBcdSQXIRt6OHdvoTpqeTi8nTzL6BgAAAAAABhNBUoHMNsbems30ttXVjXCpXGb8DQAAAAAADA6CpIJlZ3qbnZXW1qRabSNcarXS8be77iq6SgAAAAAAAIKkgdV7xreZmWJrAQAAAAAAkAiShkKjwXgbAAAAAAAoHkHSgOtdyM3OJAAAAAAAUKSxC5KazY3F1sMgSTaus51Ji4sESgAAAAAA4OiNXZAUx+llmARBupC7Wt1YxE2gBAAAAAAAjtpE0QVg/7JAqdFIv5+fL7YeAAAAAAAw2vrakWRmzzSzT5vZVTN7+Q7H/ISZfdLM7jWzpX7WM0qys7r17lCanqY7CQAAAAAA9E/fgiQzCyS9WdKzJN0u6UVmdvuWY26T9ApJ/7u7P0XShX7VM6rMNsbdepdys5gbAAAAAAActn52JJ2SdNXdP+fukaS3SnrelmN+WtKb3f3rkuTuX+5jPSOrUtm8P6nV2ryYOwuW6FgCAAAAAAAH0c8g6bGSrvV8/4XObb2eLOnJZvanZvZBM3vmdk9kZvNmdsXMriSJ96nc0ZAt5g6C9PveYCkMpYsXCZMAAAAAAMD+FH3WtglJt0mak/QiSb9uZtNbD3L3S+5+h7vfceyYHXGJw6v3bG/ZJQylJTZRAQAAAACAfehnkPRFSY/v+f5xndt6fUHS3e7edvf/Kek+pcES+iCK0ut77mF/EgAAAAAA2Lt+BkkflnSbmT3RzEqSXijp7i3H/L7SbiSZ2S1KR90+18eaxloWJCVJuj/prruKrQcAAAAAAAyXvgVJ7n5d0nlJ75X0KUlvd/d7zey1ZvbczmHvlXS/mX1S0p9Iepm739+vmpCOu2X7k65eZQE3AAAAAADIr687ktz93e7+ZHc/4e6v69z2Kne/u/O1u/svuPvt7v433f2t/awHmyVJujNpcVEqlTbO7MbYGwAAAAAA2M5E0QWgOFlnUhynoVK7nY6/NRrp7fPzxdUGAAAAAAAGT9FnbcMAyMbdgkA6cUJ66KF0IXdvlxIdSgAAAAAAgCAJm6yupp1JWYdSq5WOvy0sMPYGAAAAAMC4I0jCDXo7lLKLexooNRrS0lLRFQIAAAAAgCIQJCG3ycmNHUp0JQEAAAAAMH4IkpBbFG2Mui0uMuYGAAAAAMC4IUgqWBRJzaa0slJ0JflkZ3pLEqlel+66q9h6AAAAAADA0SFIKlgUSXEszcwUXUl+2d4kSbp6lbO6AQAAAAAwLgiScCBJsjHqVi6nodLx4+nXjL4BAAAAADBaJoouAMMt60yK43R/UjaqF8fp6Nvycnr//HxxNQIAAAAAgMNBRxIORRBI1ao0OytVKhu3x/FGt1KplHYs0akEAAAAAMBwIkhC32S7lJIk7VZqt9MxuHo9DZfYrQQAAAAAwHAhSELfZd1KvUu6s91KCwvpbeXyxo4lwiUAAAAAAAbT2ARJ7unuHhQvC5SCIP29ZB1LrdaNi7sJlQAAAAAAGBxjEyQlSbqvB4Mn61jKupayYCkMpXPn2KkEAAAAAMCgGJsgCcOjdxQuO/vbXXcVXRUAAAAAACBIwkDLdipdvcqoGwAAAAAARSNIGjFxPHq7oLJxt637kxh5AwAAAADgaBEkjaBR3AWVLefO9ietr6cjbwsLabBEqAQAAAAAQP9NFF0AsBdBIFUq6dfNZhqatVppqNRopLfPzxdXHwAAAAAAo4yOJByaKErDnCg6mterVDZ2KB07lo6+XbxIZxIAAAAAAP1CkIRDUyqlYc5RBUm9zNLrbI8So24AAAAAABw+giSMjKw7KUnSUbdsMXe5LJ08SbgEAAAAAMBBESRhpGRLuaWNxdytlnTffZuXc09PEyoBAAAAALBXBEkYSUEgVasblyxcck+DpWwEbno67VSiWwkAAAAAgN0RJGFs9HYrSWnHUrMpraxsHoWbnt4ImAiXAAAAAADYQJA04qIoDUuKWIA9yLKOpdlZaWYmvS0bhcver3pdOneOQAkAAAAAgAxB0oiLIimO0zOqYWdZt1IWLp06ld4exwRKAAAAAABkCJKAHfSOwmWBUrZXiUAJAAAAADCOCJJQOPfBHr/beia4MEzP/jYxQZcSAAAAAGC8ECShcGZpx8+gBkmZLFAKgjT86u1SKpfT8cFsWTfhEgAAAABgFBEkYSA1m2lQM8iyUClb0t1up9dhuPkMcHQtAQAAAABGBUHSIRn08axhM+ghUqZSSRd0V6ubO5aycGl9Pe1aOn8+DZbKZYIlAAAAAMDwIkg6JEnC2dGwITsD3NRUep0FS61WGiwtLKSflSCgawkAAAAAMDwIkoA+q9XSzqXepd3u6Shckmx0LZ07lwZKhEoAAAAAgEE1UXQBwLjJwqQ4Tr+uVNKxSElaWZHuuy+9np8vrkYAAAAAALZDRxIwACoVaXZ24/urVzkDHAAAAABg8BAkAQNkdTW9znYqhWG6T2ligpE3AAAAAEDxCJKAAZMt6s7OBOeejsHV69LZsxvLuQmWAAAAAABHjSAJGALZXiVJiiKp0UiDpQsX0hG4cplgCQAAAADQfwRJQIHc00XbDz2UhkNRtPOxWafSqVPp96dPS6VSOgLXaqXB0sWLhEkAAAAAgP4hSMLYW1lJR8fcj/61zdLXTpJ0H9LNgqRetZp0+XJ6nY3AHTu2sVOpXE5DpiDYWNqdjcQRNAEAAAAA9osgCWMvW3BtVmwdB2W2sVOp1ZLa7Y2l3c1mGlLV69K5cwRKAAAAAID9mSi6AKCfoii9uA9/UCRt7EeStg+CKpU0NNp6W62Wfr28nF7X69I990jnz0vXr6fvz7Fjm9+jSuXG58+eZydnzkjz8/l+FgAAAADA8CFIwkiLonR0rHdZ9TCLonRkTZKWlvb++CxUajTSMbgk2biv92spvX+rLIjaThyn91+8uPe6pN1Dqn4g+AIAAACAvSFIwtjKunvy7iU6CnlCr+PH03G8lRVpZmb/r5W9Thyne5aygKnZ3OjgSpLNu6PiePf6twug8qjX9/e4g8hG/UbZdp1lGB5FBKwAUvzPBgAAtkeQNCa2jjsh7ewJw+HrVlpdTevu139g9gYPzebu4dGwG/Wfj3/2h1s2ygrgaIVh+s/ffrp/AQAYNnttUiBIGhOjNN6Fvck6r44fz398FEknTqTfbw2TzNJ9SpVK+pwH6YoCAAAAABRrdVW67778xxMkASMu67zKK9srNTOzcUa7INgIk9w3RtjW16Vr19LbT51iDAAAAAAAhs3c3N5WjRzrWyUARk4QbFykdIdSq5Ve6nVpYUEql9PwqlSSpqfTP5S2O8McAAAAAGD4ECQB2JcsUKpWN4Il9zRUarfTSximAdPiYhoqESgBAAAAwHAjSOqDKNo48xUwSPp5prrebqUsWOrtXArDNFCiQwkAAAAAhhdBUh+USukOmSQpuhJgIzSKoo19Sf0IknbSGy4lycYIHGNvAAAAADB8CJKAEZeFRqVSsXX0ct889lYqpbuVCJYAAAAAYLARJAE4Utst7G63exd211Uul1UqlTQ9Pa1LJEsAAAAAMDAIkgAUZnOoFEuqy31OrVZL7XZbYRjq3LlzmpubI1ACAAAAgAFAkASMCJerGV+Xa3i3vAdBsOkiSXEcq16va3FxUdPT04RKAAAAAFAggiRgRNhkrPhEU8nw5kjbygKlJEm0vr7eDZXK5TKhEgAAAAAcMYIkAAMv61CampqSlIZKrVarGyqVSiUFQaByucxeJQAAAADoI4IkYIg1P9pU3IyLLuNIbR19S5JE7Xa7Gy6FYdjtWCqVSpqYmKBzCQAAAAAOCUESMMTeuHZFJ+L1ossoTBAEqlar24ZL2cLubMfShQsXuuFSuVzuXgiZAAAAACC/iaILAIDDFASBKpWKJKnZbEpS9/swDCVJ7Xa7e3y9Xtfy8rIkaX5+/ihLBQAAAIChQ0fSgGg0GnRFAH1Sq9VUq9U2dS9lX0vpmeEWFhY27VmiUwkAAAAAbkSQNACCoKkwfJqWlpaKLgVDpNFo6F16V9FlDIWVlRU1Gg1FUXTDfVmw5O6b9izV63WdO3eOQAkAAAAAehAkAUOoWm0oDJ+m9+l9RZcyFFZXVxWGoUql0k2PyzqUMtl+JQIlAAAAAEgRJAHANnoXeLu76vW6FhYWGHsDAAAAMNYIkoAh9ll9tugSRl6lUtHU1FQ3UMrG3hYXFwmUAAAAAIydvgZJZvZMM/u0mV01s5dvc/+LzewrZtboXP55P+sBRk1TzaJLGBuVSqXboSRJSZJ0u5SyBd2ESgAAAABGXd+CJDMLJL1Z0rMk3S7pRWZ2+zaHvs3da53Lb/SrHgA4LL27lNy9u6B7cXFR5XJZ5XKZbiUAAAAAI8ncvT9PbPYMSa929x/pfP8KSXL3f9NzzIsl3eHu5/M+71NvMf+9ZzPwDc4fAAAgAElEQVSRBwAAAAAAcFhOviX5iLvfsdtx/UxkHivpWs/3X+jcttWPm9nHzOwdZvb4PtYDAAAAAACAAyi6teddkm519++R9N8lvWW7g8xs3syumNmVI60OAAAAAAAAXf0Mkr4oqbfD6HGd27rc/X53f7jz7W9Ievp2T+Tul9z9jjwtVgAAAAAAAOiPiT4+94cl3WZmT1QaIL1Q0pneA8xsxt1XOt8+V9KndnvSv1x7um7/LxuNSXEs9ey97apUpGbPCa3iePP92WO2Hnez593ptXqfv1pNr8Nw8+vc/DmaiuMpnT59WtJlLS/feHyl0jmyeeNzbL1v6+vudNt2P8tTn3pZV6/W9I1vTO9a93Z1hGF623b33ex93nr8djVv/Xml7d+Pra+53XNn7+nWeisVqVZLj+v9PexUd6UiPelJDV2/vqaPf/wHle6Y31xnrSY1Gjf+/Ft/9u7PXLmu+ERTwWcrUmtih993Q2H4NEnS9+rr+kQwLU2mj9PHqwrMbnit3s/m1vfj0ZOJbis9qM9Ej9KD0THFsXT69Oa68/6ztvXn2+22jftiBcGfSZrbclvve1pRc8uL7XSbpO7tlUpFYRiqWq1uuq33cdlrxZ1fRPa6lZ4P3HbHb/0Zeh9748+3+WfZqtlsdp9Dko4dO6YgCDQxMaFSqaQ777xT8/PzNzwOAAAAAPZrbk6q1yXJdjky1beOJHe/Lum8pPcqDYje7u73mtlrzey5ncN+zszuNbOPSvo5SS/uVz1AETxK1Gysy6Ok6FJuKokSvTH8iG4rPVh0KVAaRAVBoCRJ1G63u2eFW1hY4KxwAAAAAArV1x1J7v5ud3+yu59w99d1bnuVu9/d+foV7v4Ud/9ed/877v6X/awHOGq3lR7UG8OPKBnwIAmDKQuUsou7q9VqqdVqqV6va2FhQaVSSeVyWdPT0wRLAAAAAPqu6GXbAIA9yEIlSXL3TR1Li4uL3VDp5MmT3ZCJDiYAAAAAh6WfO5IAAH2ShUm9u5fiOFar1VIURZKkdrutdrstSarX61ruLP5izxIAAACA/SJIAoAREQSBKpWKap2N8cs9G+PjOFYcxzp37pwkwiQAAAAA+8NoGwDcRBRFajab3S6fYZWNxMVxrMXFxU27lRh7AwAAAJAXQdIhi+Om3B8uugwASncIxXF8oBAoiiLFcaxSqXSIlRWr92xw6+vrqtfrm8KlUqnE8m4AAAAA2yJIOkSVSlMnTlxRkrSLLgXYl7g5WmeXMzNJOpRuomazeeDnGBS9Z4KbmppSEASbwqV2u71peTcdSwAAAAAyBEkAujz2oksYWHEcF11C31QqFVWr1U0Bk5R2LrVaLdXrdS0sLKhcLisIAk1MTDASBwAAAIwplm0DALbVezY4d1er1ereF4ah6vW6Go2GJJZ3AwAAAOOCjiQAwE1t7VTKbpPSQOncuXM6fvw4XUoAAADAGCBIAgDsS2+wdP/993e7lBYXF1nWDQAAAIwogiQAwL4FQaDZ2VlVKpVNu5V6l3VPT0+rXC6rXC539ytl3UsnT57sHjM3N0dHEwAAADDgCJIAAIemdwwuW9YdhqEefvhhtVqtG5aWr66uqtVqaX19XY1Go9vRVC6XVSqVVCqVukETARMAAABQPIIkAEDfBEGgarWqqakpVatVnT59Wmtra1pbW9Ply5dVq9W690tStVrtBlDtdlvtdrsbRi0uLqpUKhEqAQAAAAUiSAIA9E2lUlGtVst1bK1W6wZL1Wq1OyrX2+HUbre7oRKBEgAAAHD0CJIAAEOh98xxvXuYGH8DAAAAjg5BEgAUrNlsqtlsFl3GUNnapZSNvy0sLGhiYqK7xJs9SwAAAMDhIkgCgILFcXzDEmrk07vcOwgCubviOO4u8d5pz1IWLpXLZc4UBwAAAOwBQRIAoHBRFKnRaGhlZeXAz1WtVru7lnbas5SFS61WS/V6XQsLC91gaXp6mnAJAAAA2MFE0QUAAFAqlRSGYd9fJwuW4jje9LW7q9VqSZIefvhh1et1vf/979fi4qLMTJVKRZJuWBx+5swZzc/P971uAAAAYFAQJAEAxloWKEnqBkbNZrMbMElpx1TWNZUJw1CNRkNLS0uSpJWVFV27dk3Xr19XuVzWnXfeScgEAACAkUOQBADomyx8iaJIpVKp6HJyq1Qq3QXos7OzkqRGo3FDR1KvbC+TpO4+posXLxIoAQAAYKQQJAEA+iaKIrVaLVWr1aJLObBarabLly/veP/c3JwajUa3mylJkm6gdOHCBZ06dYpROAAAAAw9lm0PkHTR7JeKLgMA0GNlZWVfi8B7F3xnS73Pnj3bXerdu9z75MmTLPkGAADAUCBIGhDValVhGGp19cvb3u9Rorh5Xers6wCwf3Ecd8eWgN2srq4qDEPNzMzs6/FZoJTJzhaXXZrNplZXV7W+vq56va4LFy6oVCqpVCoRLgEAAGDgMNo2JJIo0Ym4qc8GFTX5tQEH1nvWLuCoBEHQ/extPRNc7yLv3l1L9Xpd9XpdCwsL3eXfe3m9vLJ6DtPNdkoNEkYOAQAA8iORAIAx5e5qNpvd7hcULwtelpeXu4FTHMeStOcQSVL3sXmEYbjn599NvV4/9Ofsh3q9rnPnzhVdxkjoRyAJHNSwhNoAUISVlRWtrr5NUv4/KwmSAGBMmZniOB66M6qNm96uot5Ouu1ComPHjsnMjqw2AAAADLd0jcPanh5DkAQAwBDaLlAyM83OzjKqBQAAgFzm5ua01yZylm0DADDkekOler2uxcVFTU9Ps6QbAAAAh44gCQDQd5wlr/+ys8MFQaAkSRSGoRYWFjQxMaG5uTnO/gYAAIBDQZAEAOi7vSx9xsFUKhVVq1UFQSB3VxzHWllZ6Z75rVwub7rQuQQAAIC9IEgCsK1Gc12RJ0WXAexZFEVqNBqKoqjoUgZCtVrVzMxMN1hqtVqbLmEY6uLFi4RJAAAAyIUgCcANgsp1hW/8iKKEIAnDJ4oihWHImei2qFQq3dG3arXavUxOTioMw+5eJcbgAAAAcDOctQ0AgDGWdW4lSaL19XU1Gg2FYah77rlHFy5c6B5XKpVUq9UkibPCAQAAjDE6kgAAGHNZl9LU1JSkdBwuSZJNI3BZyFSv13X+/Plu9xKdSwAAAOOFIAkAMFCazSZneStQrVaTmd0wApeFTKdPn+6eFa5er28aiSNUAgAAGH0ESQCAwmXjVVEUKY5jzvJWsLW1tRt+B9nv6MyZM919S9LGSFwWKpVKpe7Z4AiXAAAARg9BEgCgcFlIwYLswVUqlRSGoZaWlrq3Zcu7s26lJEnUbre7o3D1el0LCwuESgAAACOEIAkAxlwcx4yS4cCyUKk3XAqCQO7eHYM7d+6cTp482e1aonMJAABg+BAkAQAYJRshg7RjKhuByy6StLq62u1aarVaeuCBB1Sv13X27NlN4VJ2PT09rZMnTxI6AQAADAiCJAAARsig7piqVCqanZ1VrVbrBkvValXHjm38VSSKIkVRpGaz2b2W0vCpd1wu28U0MTHRDZcImAAAAI7GRNEFAACA8RYEgeI41uzs7A33Xb58WXNzc2o0GpLSjqs4jpUkiaQ0fGo0GgrDUI1GQ0tLSzpz5ozm5+eP9GcAAAAYF3QkAQCAgbKysqJGo6GVlZUb7qtUKqpWq92OplOnTkmSTp8+rePHj3d3MdGhBAAA0B8ESQAADICss2a78GTcrK6uKgxDzczM5Dq+Vqvp8uXL3ePjON40AsduJQAAgMNDkAQAwAAolUoKw1Crq6tFlzIwGo3GnsOf3sXeSZKo3W53zxp39uxZdioBAAAcEDuSAADAwKlWqwrDUEtLS/t6fBYmSZvPSti7U+n973+/FhcXJUlmpkqlIintcJLSEbuZmRl2LgEAAPSgIwkAgBz20x2DwREEgU6fPt3dqVStVuXuSpJESZJse6a7a9eudUfkGI8DAABIESQBALCLg3bHYLDUajXVarXuGFy1WtXp06e1tramtbU1Xb58WZcvX1apVJKUjsitr68TKgEAAIggCRgbcdwsugQAGDpZ0DQ1NaUgCDaFSnfddVfR5QEAABw5giRgDDzqUWs6ceJK7uN9yzUAHIVBP2tdpVLphkpSOvp2/PhxlUollctlTU9P06UEAABGHkESgBsknQTJii0DQ8Dd1Wg0FEVR0aVgyGXjg8Ny1rrJyUm1Wi19+ctfVrvdVqvVUhiGWlxcVKlUUqlUYgQOAACMJIIkAMC+mZnCMMwVJGVny+I/qovl7mo2m30P/5rNpprN0R2pzd6/7Oxw2b6lJEnUbrfVbrcVhqHq9brOnj1LqAQAAEYGQRIA4EiUSqUjXVg9yiHGQWRnKMsWSfdLHMfbnglt0GTB2s1Czr2EYlmglH3d+xws6wYAAKOAIAkAMJIGPcSIokjNZnOgdwINgn7vTUqSpPv1TiHnQUKxLFianZ1VtVpVkiR64IEHVK/Xde7cuW6olF0IlwAAwKCbKLoAAADGURRFiuNYMzMzRZcysLK9Sf1WqVRUq9X6/jrSRpdSHMdKkqQ7/pbJRuHGXW83F1KVSqXoEjDAjurPMACj5cyZM/t6HEEShorL1YxjlY4dU8loqAOAw5TtsRpE2WjZqPzHdBAEA981BwAARle9Xlej0dDx48f3/FiCpBHQ7P5FdPT/713ikk40VZoIpKtTRZcDACMl22M1iEYxdNmu6yZJErmnp86cnJxUqVRSrVbTmTNnND8/f9QlAgCAETU3N7epK3ovaOkYAfGJdcUn1osuAwCAgbeysqLl5WV99KMfLbqUG1QqFR07tvFXsyiKtL6+3l3SXSqVFASBpqenWdgNAAAKQ5AEAAAKF0XRput+mZmZURzHWltb6+vrHEQQBDp9+rRmZ2c1NZV23yZJona7rSRJ1Gw29dBDD206C1y5XO4GTCdPnlS5XFa5XCZsAgAAhy5XkGRmTzaz95nZJzrff4+ZvbK/pQEAgFHg7rueoS4LkEql0g239TtcGnTZmd96zwBXLpclqbuwOzsLoCStrq6q1Wqp1Wpt6mbKwiaCJQAAcBB5O5J+XdIrJLUlyd0/JumF/SoKAACMjiRJ9nWGuu3CJWzIdixl3Uuzs7NaW1tTrVZTtVrt3p91M7VaLYVh2O1iolsJAADsR94gqezuH9py2/XDLgbA8InjptxHbwkuipF1VKAY4975M6q2djRlXUy9o3GESgAAIK+8QdJXzeyEJJckM/uHknbuTwcwNk6cuKLJSZa943CM4pm5hskgBEmEiUcjCAJVq1UlSdJd6L2wsKAgCDQxMUG4BAAAdpQ3SPoZSb8m6bvM7IuSLkha7FtVAABgLBEmHo1sXHByclJTU1MKgkDu3h1DbDab3XApC5Z2Wuid3U7wBADAeMgVJLn759z9hyQ9RtJ3ufusu3++r5UBAACgL6IoUhiG3UCpUql0R98mJycVx7EmJyfl7t1wb6eF3q1Wq9vVdP78ec4WBwDAiMt71rbXm9m0uz/o7utm9s1m9q/7XRwAAACOVhYulUql7gjc2tratgu9s8vU1FR3VC47W9zCwsKmjqWsa4mQCQCA4ZZ3tO1Z7r6WfePuX5f0o/0pCcMqih6WJLknBVcCAAAOaq/7qmq1miqVSvd7d9/UsdRsNtVoNFSv13Xx4kUdP35cQRAwFgcAwJDJGyQFZvZN2TdmNinpm25yPMZQFLUlXZbZQ0WXAgAADugg+6p6zxKXdSxVKhXVajWdPn1atVpN999//6Zl35xBDgCA4ZA3SPodSe8zs5ea2Usl/XdJb9ntQWb2TDP7tJldNbOX3+S4HzczN7M7ctYDAACAIVKr1XT58uXuJdvLlC373hoqlctllUql7vXExMQNC78JngAAOHp5l22/QdLrJH135/Kv3P3Omz3GzAJJb5b0LEm3S3qRmd2+zXFTkn5e0p/vrXQAAACMgkql0u1cykKlVquldrvdvY7j+IaF31nwtLCwQKgEAMARmch7oLu/R9J79vDcpyRddffPSZKZvVXS8yR9cstx/0rSGyS9bA/PDQAAcCiiKFKj0ZAkXbp0SfPz8wVXtHcrKytaXV0diRCld89S756mUqnUXQR+5513amlpSY1GQ81mU3EcKwxD1ev17u9yGH+PAAAMg7xnbXuBmX3GzEIze8DM1s3sgV0e9lhJ13q+/0Lntt7n/T5Jj3f3P9zl9efN7IqZXUkSz1MyAABALlEUdb9eWloqsJL9m5mZURiGQ1t/Hjv9nrIRuSAIJElhGOrixYsjEaoBADCI8nYk3SnpOe7+qcN6YTM7JulXJL14t2Pd/ZKkS5I0MXEHSRIADKHsPwJXVlYKrgR5RVGkKIo2dYKMqlqtVnQJyGG331M2FpeFSRKdSQAAHLa8y7ZX9xEifVHS43u+f1zntsyUpKdKumxmn5f0A5LuZuE2AIymLEiamZkpuBLkFUWR4jge+RAJ/ZXtNWo0Gkeyw8g9/X+OYRh2l3azPwkAgMOTN0i6YmZvM7MXdcbcXmBmL9jlMR+WdJuZPdHMSpJeKOnu7E53D939Fne/1d1vlfRBSc919yv7+UEAAACwf+7eXWh9mLLnq9VqajQam8bS3F3NZnNPr7mysqJGo3HT7sZs1C1b2t1sNrtLuQmWAAA4mLxB0qMlPSTphyU9p3N59s0e4O7XJZ2X9F5Jn5L0dne/18xea2bP3X/JAAAA6JWFKgcZHTUzSTr0IElK9xhdvnz5htE0M9tzeJXtg1pdXd31NavVqqrVqmZnZ1WtVuXuarVa3bO9XbhwgUAJAIA9yrUjyd1fsp8nd/d3S3r3lttetcOxc/t5DQAAgHHWaDQYHd2DIAhUqVQUhqFarZbq9bqWl5e1tLSkM2fOsFMJAIBd5AqSzOyRkl4q6SmSHpnd7u4/1ae6AAAAsItqtaowDLtnLBsnvcvg96NararZbCqOY8Vx3A2UJBZ0AwBwM3lH2/5fSd8m6Uck1ZUuzl7vV1EAAAAYHVEUaW5uTo1GQ41G41BGyUql0qHtdMp2KsVxrPPnz6tUKnV3KTH2BgDAZnmDpCe5+7+U9KC7v0XSj0n6/v6VdTAeu+JmXHQZAAAAR2q3JdRFiaJIjUajuyOpd+H2oMh2KrXbbbXbbbVarU1nfuu9TE9P6+TJkyztBgCMpbxBUrtzvWZmT5VUlfSt/Snp4CYV60RMwxQAADg8ez272FELgiDXEuqi1Gq1bRduZ6IoGoj3OBsT3Hrmt95Ls9nU6upqd2l3FjZNTEx0wyUCJgDAqMobJF0ys2+W9EpJd0v6pKQ39K0qAACAARPH8b738WB3hzmqdpiCIOie/a33LHC1Wk1TU1OS1A2b4jhWs9lUo9FQvV7XuXPnNnUv0cEEABgFeYOk97n71939/e7+ne7+rZL+qJ+FAQAAAIMk65rqHR/MOpeykKlSqahWq6larXaPWV1d7XZb9XYw9YZLBEwAgGGRN0j6b9vc9o7DLAQADoO7D8RoBIDhsrKycqiLoDGaoihSHMeamZnZ8ZjeEb7Z2Vmtra11vz516tSmcbns31dZB9Pi4uKm8Tg+iwCAQXTTIMnMvsvMflxS1cxe0HN5saRHHkmFALAHZjaQoxEYbtnniZBhdPXuFRrERdAYLpcuXdKHPvQhLS8v3/BnRrbUOxuRO3XqlCSpWq0qSZLuv8Oy0TgCJQDAoJnY5f6Tkp4taVrSc3puX5f00/0qCgCAQRJFUXcB79LSkubn5wuuCP2w0xJoYK+Wlpa63Ut5gsnss9d7Zrvl5WUlSaJ6va5GoyFJ/NkDABgIN+1Icvc/cPeXSHq2u7+k5/Jz7v6BI6oRAIDCZXtPgH7IxnLdvehS+urSpUtqNBqbdgyNqkqlcuDHHzuW/lU9DEOdPXt20+JuOpUAAEXJuyPpH5jZo83sEWb2PjP7ipn9ZF8rAwAAYylbaDzqoUqvbKTJzIoupa+WlpYUhuFNdwxhs6wbUpKazaZWV1f1wAMPqF6v6+zZsyztBgAcubxB0g+7+wNKx9w+L+lJkl7Wr6IAYJSxEBy4uWwkaNRDFeyNu49NN9NW2ZnhZmdnVavVup1KUvrPy/r6uprNZndp98WLFzU3N9ftXiJYAgAcprxB0iM61z8m6b+6e9inegCgb5rNZtElSGIhOIDhk42kFblw3swUhuGmxejjLAgCnT59WidOnJAkxXGs48eP6/Tp090x3NXVVYVhqMXFRcbhAACHZrdl25l3mdlfSmpJWjSzx0j6Rv/KAoDDF8fxphEBAEer0Wjo+PHjRZeBfehdGL2fs9olSaLl5eVu6DHsBumzvLq6qiRJJEkzMzO6fPly9765uTktLy8rjuPuOByLuwEAB5WrI8ndXy7pb0m6w93bkh6U9Lx+FgYAAA4uiiI1Go3CO+CCIKCbZMjVarV9L5x3d8VxPBK//8P8LK+srBzKuN5u/5MkCIJNi7sXFxdVLpe7+5VY4A0A2IubdiSZ2d919z82sxf03NZ7yO/1qzAAAMbRYY9gRlGkVqularVaeJgEYLNs9OyozgiZBU5xHKvVaklS98+F9fV11et13XPPPVpaWtKZM2foWgIAbGu30ba/LemPJT1HkkuyLdcESQAAHCJGMDGuskBjnM7WV4RKpbLp+yzEajQaajabiuO4GyhduHBBknTq1CmCJQBA125B0rqZ/YKkT2gjQFLnawAARl72H1bZiNig7EUBihZFkaIoUqlUUqlUOpTnk27ofschys4autPvrFKpdP/MS5Kk27WUBUsXL16UtBE+ES4BwHjaLUjK/pfFSUn/m6Q/UBomPUfSh/pYFwAAAyGOY0kbI2JHNYKCo7eystLde8OemN2VSiW1Wq1umITBt/WsoTsF5EEQdDuXeoOlZrOpSqWilZUV3Xfffd2upVKp1H3O69evK0kSVSoV1Wo1wiYAGEE3Xbbt7q9x99dIepyk73P3X3L3X5T0dEnfcRQFAsDNZH/BBYCD6l2evJ8zkwFHyd0PtEi/93G7LQ4/duxYN1y68847NTMzoyAIul1LzWazG7a3223Fcaxms6l6va5z587p5MmTKpVKmpiYYKE3AIyAXGdtk3RcUu+/paLObQBQKEIkAIepVqvp+PHjhZ7pLgsIDnomr0F02Mvkx5mZKQzDA31Os66h3SRJ0v06C1krlYqq1aqq1apmZ2c1OzurarWqIAi6odPk5KTiONa1a9e6AVO9Xtfi4mL3LHEESwAwfPIGSb8t6UNm9moze7WkP5f0W/0qCgAADI5x+4//mZmZA/8H+kFkAcFhnF5+kARBMLLhf9aBs5OVlZWhDwfzhk6ZWq2mU6dOSUrHILOAKetkWl9f///Zu/8oyc76vvOfp++opB71UC1AGhoEAmQhLzamCFg29jhq2zgGs4A3i20Y/147A5qQxOs9Z+JNfOwsmxwnxJP4nD1s4s7acfZHL7G9m2M2BuNfTGfaGAaBS0YgEBoJGKGihITmqkvTM3d069k/nudb93arR1PdXVW3frxf58yZ7uqqe5/73Od57r3fen6o2Wz2AksHDx5UrVbTwYMHtbi4SHAJAMZYX4Ek7/0/k/Szkp6I/37We/9rw0wYAAAYD9P68A8M0rPVEwtOTltwcC+sJ9OhQ4ckSfV6vTdE7vLly9rc3FSaplt6LRFUAoDxcrXJtnu895+W9OkhpgXoWyfPVfNd9d+pDgAA7Ed5MnJJlfXYwvSwHk7NZlPS1t6PeZ5rY2Ojt2KcTepd/hwTeQNANXgKx8RJnJR7KSuN1wcAAMNlQaRGo9GbXBnV8d5v+X/aJEmiQ4cOPWNSbykENbfPtUSvJQAYnb57JAEAAGC2NRoNnTp1SouLizM3d9aoWM+vqwXqbAJs59wokrUrNmH84cP7X5tnYWGh97OVv+XlZZ09e1Z5nitNU62trfV6NdFDCQCGjx5JAABg6nnvdebMGZ05c6bqpGBALNAybT1y2u220jTtDeOaRP1MGJ9l2b4mIF9YWFCSJJqbm9P8/LzSNNXx48d7PZRuv/323iTfNoE3PZcAYDAIJAEAgKnnnNPm5qY2NzerTgoGxAJJ49QjZ7/BkXHS6XSG2uusVqsNZAJy55yyLFOSJFt6L7XbbV2+fLk3LM7mW3r3u99NUAkA9omhbQAAAMAAWHCkbFBDvEbNVqFLkkRSESQ7fPiwlpaWqkzaMywsLPSGvZnl5WWtr6/3/i5JaZrKe89wOADYJ3okAQAAAENQr9cH0utmWCw41M/E6VmWKU3TsQsiee/V6XTUbDav2sPIgmJJkuiaa67pDYejdxIA7A6BJAAAgBIbljRtc+9g+vUTTCmz4NAkz8fU7XZ7vYxOnjy5q89JoRdZeQU4m1uJ4W8AcGUEkgAA2IXdPqhh8ljvkXGaewe4Guv9tLq6WnVSRs7q6m57S9Xrdd1xxx1KkkTdbledTqc3t5INf7vrrruYrBsAtiGQBABAn2b5QQ2omvUQ62cY1qSZ5Am6J3mCcUt7rVZTvV7XkSNH1Gg0lCRJ79/2ybqPHz+uw4cP6+DBg6rVajpw4ABBJgAzh8m2AQAAsC+7mWtnr8pDkaZp9T0LUE+qnSYYnxRZlmlzc1P1en3Hv5dXgZPUW8Xu8ccf701Gbq+vra3pzJkzOnHihA4fPqxz587pxS9+saTQU+ro0aNM6g1gahBIAgAAwL6laaq5OTq7Y3rZ6nDNZrMXVLrS65ubm7r//vslSffffz8rxAGYKjN/te/kuXIm0wQAANgz64k0inmlbF9Mhj49Op3OlsDMuLpSz7tGo6FTp06p0Wj0hsfV63XV6/XeSnFpmvbmW7KhcQyJAzCpZr5HUn7rhvTAgnRx5rMCAHAV9vBgE24P4pvlRx55RI8++uhUzvuC0SrPITTJq3BdzSiDVjuZlXweJRsmZkGXnVgQp8qgiw3je7Z0Phubb8mGZm5ubmptbU1ra2t617vetbGbNzAAACAASURBVOW9O+1j+1C7nTQajT2lbTuG4gF4NkRPAADoU5ZlvZv71dXVgdxkP/roo0rTVPV6farmfcHolecQwvCQz9WwIM6kLnaQJEkvENTpdLbMsbSTnf7eT68tG0K3H2maqtlsDiSvW62W2u22sixTlmXy3vfq0KjsNfA3CP0E/6bFoIKYGI177rlnX58nkAQAwC7YXBjArFpZWVGn02Fo2T7YQ3WtVpv4gJT1EKOH1u6Vg0tWJvI815133llxygan3W73AmBXC54BGJ3z58/v6/MEkgAAwFibhLlTZsnq6qryPK/0W/5JZwGDSQy8NJtNHT58uPe7BZKGdSwWYLGJ3FutlpaWloayrypZmajX61M1rGx5eXnH10+dOnXF96+vr28Z7njkyJGpyhNgHOx3ePjMT7YNAADGG99iA1dnAZ1WqzW0fdTrdaVpqna7PbR9bGcBFhsONY1BJJMkyUQP37uaVqulZrN51bmuFhYWehOV53mutbU1nThxgknJgTFCIAkAAABXZZMdDzNQgb2zQNI0B1ow2coByH6DZUmSaH5+vrfqHSvdAeOBQBIAAACuyiY7HmVvFOzPPffcM7Y9+spzK82yTqczU8N3G43GrucZrNVqSpJE3W5XGxsbWltb6wWVCCgB+7eXNohAEgAAADCF9juZ6jANe26lYRpkcC7P8972bAL7qw39mkU23O3QoUO9oJL1Ujp48KAOHDjQ6610++2303MJ2IW9tGkEkgAAACaUPYyzgtruZFlWycpzs9TzBLvX7XZ7k9hP6zxJg7CwsNDLp263q83NzS0Pwu12e0vPJQJKwOARSAIAYBdsnhi+McY4sEDSfldfmTU2gfOo821ch5lhfFgPLeYiu7okSVSv11Wv13XnnXfq/PnzOnXqlBqNhg4dOiQpBJrW1tb07ne/WwcPHmQ4HDAgBJIAAIj6mauiPJ8H3xgDe9NsNmd+bhwE9NLaKssypWk685OmW7B3P70GrdeSFHptbm5uKk1Tvfvd794yFI7AErB7B6pOAAAA46Lf3gK7nSgUQMGWOC8/5E2L8gTSVc39s76+ruXlZR09erSS/e8WvbSwk0H2ttze1ti8VE8++aTW1ta0tram48eP9/6+sLCw5f1Xuua3Wi212201Gg0dPXpUx44d23dagUlBjyQAAABgAHY7gXSn0xl4z6w8z9VsNofSY5KV1kaH4dPDlSTJlt5OewloludiOnny5CCTB4w9eiSNieLCfKnilAAAAGDYkiRRnucDD8rU6/Wh9Zp8tkBZp9N5xusEnvZmfn5eaZrqxIkTkkRPlyFJkkS1Wq1Xbt/3vvftKq+Xl5d15swZbW5uqt1ua2VlhXOFmUGPpDFRXJivrTglADBcfMuKcWSreDHBLbB7VwqK7baH1iDdc889EztsLssyVm8r6Wf+wr3a77yHWZZpbm5OaZrq+PHjzLuEmUGPJADAyNjcKKurq3xrh7FiE7suLS2p3W5XnRwA+3T+/Pmqk7AvCwsLzMcXDTsgePjwYbXb7d4XXbu9P3HO9QJ/WZb15l1617ve1Xu933mXJDHfEiYCgSQAAHbJJtiUNFGT2gIAgK3OnTvX+zJhr8MJy4G/tbW13uvee3W73Wf0qLKVK7f34svzXOvr6710XM1+g40ErbBXBJIAANglCyI1Gg01m82KUwMAkLTnHiXjKMsynTlzRrVabWKPaWVlRc1mU4cPH9bS0lLVybkiC+bceeedvYnq95Pf1gvpyJEjkkK53Cng02w2delSmB+32+32Xs/zXGma9rWv9fX1Padzt0GrQZqF3nblLx3LxzstwTsCSQAA7EGj0dCpU6e0vLxcdVKALXaa9BiYdvV6faqGTlsPmc3NzYk9ptXVVaVpOhFBg4WFhYFc021uLgsmScX9wnbLy8tqNpva2NiQFCZZp+2eHraqn3Ou96VjmqZbVtWc5KASgSQAAIApMayVwACM3vz8vDY3N6eqp9W0K8/NVe6R8mzn79ChQ0rTVHfccceOASdMJgsUSmEernPnzilJEnU6Ha2vrz9rj7B+g69VBqIIJAFARfI8H/gqJFmW9bqRAwCAyVVeEezkyZMEksaY3X+VlRdueLZeZRb4P336tBYXF9VoNCa6pwq2sqDQ2bNnJT1z4vWd5srqZ9qE7b2bdmMQc3sSSAKACg16JZJardb3uHoAwPjodDrPeMAAnHOSQlCCXknj60r3X/2sCGdBhG63q42NDa2trWl9fV2rq6sElKaITci+vdeZ9Vwa1RDQQc3tSSAJAAAAqNiwlzjH1dlQJO99L4AzDqZt/qdxlGXZlqFItiLrfvO73W4ryzJtbm4+64pwSZL0AslpmirP815aOOfT70rzaA3DoOb2nBvIVgAAwFgb9DBKAJg27XZbaZr2gkhZlqnT6expzjH7rPd+0MnEENjQtEaj0VuRdS9Dhso6nU5v8YN6vS5JfW0zSRLNz89LCsFNYBwRSAIAYAbQ2wE7IcAIXJmtnLaXlbTss4Po2WSBrPX1dS0vL2tlZWXf28QzWa+QU6dODWSYUZ7nvWuvBaj6lWWZ0jTV0tLSvtMBDAOBJAAAgBm1mwCjPcyyIhwwWlbnvPdaW1vTyZMnK04Rhs16stncSpherVZrpOfZ9rdfBJIAAABwRTbk49KlS5K0p94Z2D2GRaEsSRIdOnRIkuilMgO63W5vxb677rpLi4uLBJSm1NLSUm8OtFHub78IJAEAAOCKbIjFOE0+vBsWCJuknlT1el15nqvb7VadFGCgWq2Wzpw5o/X1dQIjV+GcU5Ik6na7Iw00YLRsHqxJqxMEkoAZkmXh22TvuTEFAMwGW5Z7kgJJwLSyVczyPJ+pwIgNJ9rt5NkLCwu9iboxndrttiRNXJ0gkATMkCy7LOmUnLtQdVIAYN+spwmr2kwmJoAHZpMtcz9LbEXA/QxLZL6k6ZUkSdVJ2DUCSQAAYCKxqg0mEXMfAdit+fl55kuaYnZNOH36tBYXFydidcahBpKcc290zn3BOfeAc+6Xdvj7u51zn3HONZ1z6865Vw4zPQAADNLKyoqazSa9YgD0jbmPAOyWDc1lvqTpZNeEbrerjY0Nra2t6cSJE2MdUBpaIMk5l0h6v6Q3SXqlpHfuECha9d6/ynvfkPQ+Sf9qWOkBAGDQyjdyNsYdADA8wwrcZ1mmTqcz8O0OA0OcZtckDoFCf5IkUb1e16FDhzQ/P68sy7S2tqbjx49reXl57IJKw+yRdIekB7z3D3rvM0kfkPS28hu890+Wfr1eEv18AQATpdFoqNFoVJ2MmdHpdJg0GZhRSZIoTdOhBO5rtdpEzNtVr9fpkQKtr6+PXWABz+7ChQt9n7daraYsy5QkiWq1ms6cOdMLKi0uLvaGv5X/jbosDDOQ9CJJ50q/Pxxf28I593edc2cVeiT9/SGmBwAATLg8zwkkAc+CSeinm7V/BBIC60k2S720kiRRnudqNpsEFCdIt9vd1XlbWFjQkSNHdMcdd/SCSleaqH43ZWFQ14bKJ9v23r/fe3+rpH8o6Zd3eo9z7phz7m7n3N3dLp2WAAAAgJ3M6iT0FmCZ9oBCOZBOIGFrfsxSXtTrdXpDT6C9njcLKp0/f77379SpU71/u9nmoHp0DjOQ9FVJLy79fnN87Uo+IOmHd/qD937Fe/867/3r5ubcAJMIAAAAYNLZN/bS9AcUkiTRrbfeqizLJrJn0l6Dftb7aHuv1IWFBYIqwIgdGOK2PynpNufcyxQCSO+QdLT8Bufcbd77L8Zf3yzpiwIAAACwL81ms+okDJwtkZ1lmWq12jP+PksBhXa7vSUgI0nHjh2rMkl92x706zfdtVpNm5ubDG/GzFtZWdkSMLc2YHl5ecv7jh49OrR2YWiBJO/9086590j6iKRE0m977z/rnHuvpLu99x+U9B7n3BskXZb0hKSfHlZ6AAAAgFlgEzJPG1sie6cg0ixaWFjQ4cOH1W63J25OrFkK+g2DzYV2+PDhqpOCCqyurqrZbPbq0E51adgB5mH2SJL3/kOSPrTttV8p/fwPhrl/AAAAzDbrvWC9WYBp0m63laYpQZkZk2WZNjc3tbGxoeXl5aH2PMF4ajQaOnXq1BX/vr130qBVPtk2AAAAMGidTmfLfCrOMc8mJpP3/qorVk77JOOj0Gq1ekupN5tNNZvNypZW70eSJOp2u1pbW5v6ecGwOysrK70yPKyySyAJAIB9aLVaajabzNkAjJk8z5XnedXJwBTodDqV7t+CoFe6zthQRoIJ+9Nut3vDgRqNRq+X1zivjmdzTWEy2JDEYQ9FLZfXYZXdoQ5tAwBg2tmwgnq9XnVSAABDQEByduw0XGgQQ4TKE6PbULRBsCG7p0+fZojbBLAhiaMYijrsfdAjCQAAABhz3vuhf5Pd6XRmMmiy05Lyk8CGb2L82Up1jUZjxx5O3vtnlMOdXtvOJqC3IW7Hjx8f26F42L+VlZVnDL2s6lzTIwkAAKBPnU6HFaNQCefc0FZis3I9i0EkGxa2n0CSza1jw6JGtZKWna9yb5eVlRV6pIyphYUFnTp1asceTs65Z8yDtdNrO7HhbXmey3uvtbW1oa/YheHbaWW+7au1VXme6ZEEAADQhyRJ+rqpByZJvV6f6XI9iFX9bG4de7hrt9sDSVu/sizT3Nyc0jSlR8oYy7JsS2+SQfcuTJJEhw4d0vz8vNI01cmTJwe6fYxWrVZTmqbPaE9s+OWpU6cqXa2RQBIAAAAwQJMyVCrP85kfGjWoVf3s4a6qBzvnnJIk6fVIIaA0fso9TNI01dmzZ4fSTliv2aWlpYFvGzAEkgAA2ANWawOwk0nruTaLw9mm1cLCgg4dOtQb6jTOq41Nm37re6PR0NLSUm+BjmG2E+vr6wQSJ5iVjXG9ljBHEgAAe1BerW1zc7Pq5AAAICkElKoc8jLpVlZWtgTgbB6a8txG99xzz5bP7PZh3ybdHlaPQAtor66uMk/ShLIyNa7zMtIjCQAAAACmiPWatQm40T+b0Ng0Go1nBObOnz+v8+fP97W9Tqezrzm49uP06dNaXFykDMygYbcB9EgCAAAAgCliE/SmaaoTJ05IYgWvfrRarV7eHT169Ip5tps5tWz4qA05HKVut6s0TemZNIPKk3QP4/zTIwkAAABA3yZlMvFZ12g0dOedd/YCCbi67Q/fk648Xxa9kqZDq9XS8vLylhUArzSx/k696QaFQBIAAMAYmvXVtGZRlmVjH6SZtMnEZ50tMT8NgQQrc8MeJjbMh+8qzM/P93qmTXoZmGU2VO3cuXO9oZdWVquYWJ9AEgAAwBhiNa3ZU6vVCNJE3vtK55aZFu12u9crZdJ72Fi92D6sLMsyNZvNXtAMW2VZNjVlYJbZIi+1Wk2NRkOnTp3q/asi8EkgCQAAAMBY6Xa7yvN8V3PRYGfjsIpbq9UaWm+7LMuUpqmWlpYGvu1pYSt/EWzDoBBIAgAAADCVrGfTrPXysl46WZb1fq5yeFu73Vae52O7lPm0s2Bbu91meBsGgkASAAAAgKnknBvpcMF77rlnLOY3q9VqStO0F0gyDG2aXfV6nYnXp8zKysqzTrpt8yoNI4hMIAkAAADAxBjnHkbnz58fy/nNDh8+LGk6Jt0eNnv4nrZhYFZn1tfXr7jKF0ZjUPVwdXVVzWbzipNuD3MVQgJJAADswzh88wwAs2IcV42zoWPj/GA+bcvaD5NNajxtcy6VV71bW1vTyZMnK07RbBp0z7DyxNs7zYU2rFUICSQBALBH9Xp9LL95BgCMjg0ju9qDYZZlla5Ed+2116rT6Yx90AvDkySJDh06JEk6e/YsPZOwZwSSAAAAIpucFgAGLcuySleie/zxx3tfftArabZZz77tQ6Ew2UY5LPPA0PcAAMCEIZAwm5IkUZqmSpKk6qQAwFDU6/WhDHMZB/S02p1pLgvTrtVq9SbZlqTl5WUdPXpUS0tLuv/++9XpdLSwsDDUNBBIAgAgGse5NwAAGJQsy3TmzBllWabFxUU1Gg0dPXpUx44dqzpp+8KKZJgl7XZb7Xa7Fwi0gJIZxbQLDG0DAIwc3xoCwOzodDrMJzcmsizrfVmSZZnW1tZ01113aXFxkevyNtyrYBis/u23fNkk20ePHpWkka8ySCAJADBSfGsIALOFINJ4WVhY0JEjR3TttddKkrrd7siuy5Oy0uks3KvYaoOjDkDMuizLekPoB1G+VldXK1llkEASAAAVmpSbagDAdDl//rykEDQZFYKK4yPLMqVpqrNnz9LzasQWFhYGPj/VqIOCBJIAAKgQN9UAgFljvWEYPlYtmxuS4Y2TzXrQtdvt3mtZlqnT6ei+++4bSpCJQBJQ4rOuOs0NZa1LVScFAAAAGJlRBnVsrqY0TXXixAkCGBWz4Y2ci8lggdhnO1dZlinPc126dGkoQ98IJAElt9We0q+nn1LWZsUmAAAwGAxhxbirYk6gWq3WG1Y3zXMRTYIkSTQ/P08waULUarXK5/AikAQAAAAMEUNYgZ01Go2BzxWDvcmyTHNzc0rTVCdPnqw6OXgWtvJblROlE0gCAADAzLP5JLz3VScFGLlBLUk+LWY1P5xzkqR2uz1Txz1prHyOeqW2MgJJAACMgVkZ+jIrx4nJY/NJ2IMUMEvojbLVoJdonyRVDHPE5CGQBADAGJiVoS+zcpxAlezbamA36I2y1TCWaJ8E1n6cPn1ay8vLlAXsiEASAAAAMEWqDiTZ8ECGCVbHyoD3Xt77vodt0hulYCtjzdrwNis73W5Xa2trlAXsiEASAAAAgIHpdruSNPJhgvbgX+UEtOPCggHOOTnnlOd577z087lJC54MY/LhckB21oIptoqbJK2vr9MzaQoMuk4TSAIAAAAw8bIsU5qmlU5AO+kmdW6gYU0+PMurylme5nmutbU15s6aYMPoaUggCQAAAJgQrCyHYZvVuYGwMwssEqCdXMPosUcgCQAA7Nlu5t4AsH+sLHd1zWaz6iRMtFmdGwiYVsPosUcgCQAA7Nlu5t4AgGGzIRzYu1meGwhXRmARZQeqTgAAAAAAYHwwtA1l8/PzStNUJ06ckCQdO3as4hShavRIAgAAAAbMhn3uRrknCIDRyrJMnU5naPXQez+xqwpmWaa5uTmlacqk25BEIAkAgF0b9s0mgMnX7XaV53lf77XJbGlTMC5ardZEBj32k+Ysy5TnuWq12oBTFTjnlKap2u22Wq3WxM0vaHOztdtthrhNqGazObDrDIEkAAB2adg3mwAAVKndbveCHpMSNEiSpJfmcddutydy4vzyELdJKRfjoNVqKc/zSgOHVj8IJAEAAAxBv71IAGCa2cTlszzhtvXMYqLpwIIQs14udsuCm5MWOHw2BJIAAAAAANim3W4ryzJ64ZTMzYUQAsG12UYgCQAAAACAHdRqNdXrdUmiF47C/G9MvD05drvoQ78IJGFKdZXnHUndqhMCAAAATBwbxkTPE6nRaKjRaFSdjLHBxNuTY1jD9QkkYSp573XrrXdrfv5C1UkBAAAAJk6WZb0VBemJM/689yNdUZaJt2cbgSQAAAAAu9bpdJicfsrZ6qT0Shp/zjnleT6yQFKWZQxxq1iz2VSr1apk3wSSAAAAAOwaQaTpVw5KzHqvJFvBjYBagSFu1UmSRGma9laEG7UDlewVAABghrRaLXU6HXnvp2r5XwDTj7mBgqWlJd1///0zH1DbrjzETZKOHTtWcYqqt7KyomazObLeYbsxqC8A6JEEAAAwZO12W3meE0QCAEwVhrg90+rqqtI07Q0NnUYEkgAAAAAAwL6cPXtWy8vLDHMboCzLRjqJer8IJAEAsEedTqfqJAAAMFC2+pf3XlIxN9Cszw9kkxpXNbnxOOt2u72fm80mw/8GKMsy5Xk+dr2bCCQBALBHTDQLAJg23W53y1Dc8mS+sxwgsHxYWlqqOCXjKUkSHTlyhDm1ZgSTbQMAAAAArojgAPrRarV07tw5ZVmmlZUVJt6eYvRIAgAAAABc0aCHt9nwuXEfJmbz04x7OsdFu93uDcU6ceLETA+FHJWqplkgkAQAAAAAuCILEAxqZS7nnPI83zJsbhxZUMSOH1e3sLCg+fl5pWlKMGkEqppmgUASAAAAxor1ArDJfoFxZgGGae+1YpP9ztocQUmSKE3T3nluNpsVp2j8WV6laTrT82pNMwJJAAAMwDguzQpMKusFYJP9AuPM2v5ZC7DMIgsq4erm5kKo4fTp01peXqZn0pQhkAQAwD4lSaI8zwkkAQAqYz35uBZhHHS73d7/a2tr9EyaMgSSAAAAAEyFWQ6i1Gq1kXypsb6+Tg8T9CVJEtXr9aqTgSEgkAQAAABgKsxyIGkUrAfu2traQCbdBjCZCCQBAPatqhUjAADAaCVJIims5EavJFxNeZJyysv0IJAEAMCMy/NcnU6n6mQAACZEvV5nRS70Jcsyzc3NKU1TerFNEQJJAACAXmUAgL7RywS7YStwnj17lvm1rsB7P1GT5RNIAgAAAAD0Lcuy3hA3eiWhH+X5tU6cOEEwaRvn3EStADzUQJJz7o3OuS845x5wzv3SDn//Refc55xzf+2c+zPn3C3DTA8AAAAAYP9qtZqk6euVxFDv4bHgI8PcJt/QAknOuUTS+yW9SdIrJb3TOffKbW/7K0mv895/m6Tfl/S+YaUHAAAAADAYWZYpyzKlaTpVPUwY6j1cTNYe6k6z2VSr1ao6KXs2zB5Jd0h6wHv/oPc+k/QBSW8rv8F7/1Hv/YX468cl3TzE9AAAAADArmVZRk+VHdRqNdXrdUl7G+JmD9TT1qsJzy5JkpmYrN3K9/bhahaAXVpaqihl+zfMQNKLJJ0r/f5wfO1Kfk7Sh4eYHgAAAADYtSzL6KlyBY1GQ41GY0+fLT9gT3tQAQXvvaTpGxa5Xa1WU5qmA5n3aNwC2WMx2bZz7ickvU7Sv7zC34855+52zt3d7frRJg4AMDStVkudTqd3QwEAAGbLfgJRmEzdbldzc3PMldQnm6h8nAwzkPRVSS8u/X5zfG0L59wbJP1jSW/13l/aaUPe+xXv/eu896+bm3NDSSwAYPTa7bbyPO8tCwsAAIDpZ/d+kzy8a5YNM5D0SUm3Oede5pyrSXqHpA+W3+Cce42k31QIIj06xLQAAMaEde+d5AkGAQDAYLRard48SdwbzJ5pHt5m97zT2PN+aIEk7/3Tkt4j6SOS7pP0u977zzrn3uuce2t827+UtCDp95xzTefcB6+wOQDAlLCLKt9AAQCAdru948+YfvPz81O36l+Z3fNOY8/7A8PcuPf+Q5I+tO21Xyn9/IZh7h8AAAAAsHdZlinLMtVqtR3/1mw2e78vLy/r6NGjOnbs2K72YXMklbeF6Zdl2Za5knZbbibdTiu6TYqhBpIAAAAAAONje/Cnn/fnea5arbbjMuZZlunIkSOSikDQrAUEsHezOldSvV5XmqZKkqTqpOzJWKzaBgAAAAAYPluSfFAWFhZ09OjR3u/NZlM33HCD7rnnnoHtA9NvfX1dy8vLUznEbTe89xOxojE9kgAAAAAAe7a6uqpms9kbora2ttb72yQP38Fo2PL2a2trarVaM92jzTmnPM+rTsZV0SMJAAAAALAn5aFyR48e1alTp3p/s+E7kxxIsp4h495DZNLNzYXQxNmzZ+mZNAEIJAEAAADAPs3qRNHlINHq6mqFKRmObrcraTpX3honls9SqEvTWJamCUPbAAAAJoT3nmEiwBhKkmSg8w5NGhvSBuxHkiS9idsx3uiRBAAAMCGccxM/TAQAgCtptVpqNptqNpsMbxtjBJIAAAAAAM9qmCtJWfCAIDna7XbvZ4a3jS8CSQAAAACAK7JVtcrz2AxSu91Wmqaq1WpD2T4my+HDhyWFACPGE4EkAAAwkeybcbq/A0C1GI6EQbLAYrvdpjyNKQJJAABgInW7XSVJIonu7wDQr2EMH2M4EgZtfn5eaZrqxIkTBJPGEIEkAAAwsRYWFlgtCAB2YVjzEDUaDdpjDEyWZXxZNMYIJAEAAAAAgLHCl0Xji0ASAAAAAAAYK1mWMffWmDpQdQIAAAAAAADKbBjm5uamTpw4IUk6duxYlUlCRCAJAABcld3MsRQvAAAYlVqtJkm9ibclgknjgKFtAADgqiyQtLS0VHFKAADALKnVaqziNmYIJAEAAADAPnjvq04CMNXKPZPuuusuLS4uElCqEIEkAJgi1mtkWEv7AgCAZ+p2u1UnAZh6SZJICvUtTVOtrq5WnKLZRSAJAKaIBZDsWxtg2pVXdGH+JgAApluSJKrX65LEam4VIpAEAAAmVrn3XbvdrjAlADCb8jxXp9OpOhmYMcyZVC0CSQAAYKI1Gg01Go2qkwEAMyvP86qTgBmTZVlvqBtD3EaPQBIAAAAAAJgoNpUDQ9tHj0ASAAAAAGBfWq0W89VhpLIsU5qmarfbDG8bsQNVJwAAAACTg7lQgGp578eyHrbbbaVpWnUyMGNsrqTjx49Lko4dO1ZximYDPZIAAADQlyRJmAsFqJhzbmzrYb1eJ5iEkbJFN/I8Z66kESKQBAAAAACoRKfTGcseVti9qs6lTbrdbDYZ4jYiBJIAAAAAAJXI83xse1hhd6o8l9Ybjl5Jo0EgCQAAAAAAjBWbD8yGr2F8MNk2AADAmOLmGQAwq7rdriSpVqtVnBJsRyAJAFCJZrPJQzJwBTapNXVkfGVZpizL5L2vOikAMPPsetlqtSpOyWwgkAQAGLkkSZSmaW9yRACYNFmWKc9z2jEAGAMWSFpaWqo4JbOBOZIAACNn3+DzTT4AAAAG5fTp0zp48KAWFxdZwW2ICCQBAEbOxrw75ypOCQAAwOgwefRwdbtdbW5uKk1T/cIv/IKWl5cJKA0BQ9sAAAAAABiBPM9Vq9W0ublZdVImRqfT6fu9SZJoYWFBkpSmqdbW1tRsNiVJNAfMtwAAIABJREFUx44dG0r6ZhE9ksbMbioJAAB7xTeiAABgEuR5rjzPd/25JEl689itrq4OOlkzjUDSmMnzp6tOAgBgyrEiGAAAmAVzc3Pa2NjQ+vo6Q9wGiKFtAAAAAABgKmRZpizL5L3X5cuXe6+fOHFCEkPcBoEeSQAAAAAAYCpkWaY8z3uLuiRJovn5eaVpquPHjzMB9wAQSAIAAAAAAFOrPJy/2WwyZ9I+EUgCAAAAAABTLUkS3XrrrZJCMIleSXvHHEkAAAAAAGDqtdttZVmmzc1N5kzaB3okAQAAYEfe+z0tuQwAo+a9lyRWJMVV1Wo1SVKapgxx2yMCSQAAANiRTVQKAOOu2+1KKoIEwLOp1+tVJ2GiEUgCAAAAAAAzw3qutVqtilMymQgkAQAAAACAqea9V6fTUZZlvUDS0tJSxamaTASSAAAAAAD70ul01Ol0qk4GcEXOOeV5vmUeLVZv2xtWbQMAAAAA7AsT82PSzM3NKU1TVm/bA3okAQCAkcvznG+uAQBAZbrdbi+YdPLkyaqTM1EIJAEAgErw7TUAAKiSrU7KXEm7QyAJAAAAAABgRGyeJu99xSnZGwJJAAAAAAAAI2KBJOsRNWkIJAEAAEwI++ZyUr/BBDB+yitYAUA/CCQBAABMiG63K2lyv8EEMD6SJJFEIAmQpPX1dS0vL2tlZaXqpEyEA1UnAAAAAAAAoApJkijPc62tranZbEqSjh07VnGqxhs9kgAAAAAAwMyyHnppmurEiRP0TLoKeiQBe+Tl1clzZd0uEVkAAAAAmGD1el1ZlvWCSRI9k66E519gj9x8rvzWjjImPAUwQ5rNJvNpAACAqZRlmebm5uiZdBUEkgAAQF+SJFGapgSSAADA1LIFLdI01erqasWpGU8EkgAAAAAAA2dfPHh68GPCzM/PSwo9semV9EwEkoCSbua3/A8AAABgbyyQZD08gEmRZVlvAm56JT0Tk20DJT7rSpLmak5i5AYATIRWq6V2u82QOwAAMDALCwtqNBpVJ2Ms0SMJAABMtHa7rTRNVavVqk4KAADA1COQBAAAAAAAgL4QSAIAALuS53nVSQAAYEedToehzhiILMvUbDbVarWqTsrYIZAEAAAmlvdenU6n6mQAAMZAkiTK85xAEgYiyzKlaaqlpaWqkzJ2CCQBQ9LpdHR/54tVJwMApppzjh5SAABgaM6cOaPl5WWtrKxUnZSxQSAJGJI8v1sb+WurTgYAAAAAYA/q9bo2Nze1tram1dXVqpMzNoYaSHLOvdE59wXn3APOuV/a4e9/0zn3aefc0865tw8zLQAAAAAAALsxPz8vSWo2m/RKioYWSHLOJZLeL+lNkl4p6Z3OuVdue9tXJP2MJEJ7AAAAAABgrGRZpiRJJIleSdGBIW77DkkPeO8flCTn3AckvU3S5+wN3vsvxb91h5gOAAAAAACAPVlYWFCj0ag6GWNjmEPbXiTpXOn3h+NrAAAAAAAAEyHLMp05c4bhbdFETLbtnDvmnLvbOXd3t+urTg4AAAAAAJgRWZZpc3NTaZoyvE3DDSR9VdKLS7/fHF/bNe/9ivf+dd77183NuYEkDgAAAACmRZZlkiTv+eIdGAabdPv06dNaXFzU8vLyzPZOGmYg6ZOSbnPOvcw5V5P0DkkfHOL+MMbyTi6fc1EDAAAAhsECSc7xxTswDFbHut2uNjY2tLa2ppMnT1acqmoMLZDkvX9a0nskfUTSfZJ+13v/Wefce51zb5Uk59y3O+celvQjkn7TOffZYaUH1bo139C88qqTAQAAAADAniRJonq9rkOHDkmS2u32TPZKGuocSd77D3nvX+G9v9V7/8/ia7/ivf9g/PmT3vubvffXe++f573/lmGmBwAAAAAAYL/q9frMzpk0EZNtAwB2p9PpVJ0EAAAAYGrZULdZXMmNQBIATKE8ZygpAAAAMCwWSErTVCdOnJipYBKBJAAAAAAAgD1IkkSSZmqIG4EkAAAAAACAPajVasqybKaGuBFIAgAAAAAA2IMsy7S5uak0TXXy5MmqkzMSBJIAAAAAAAD2yIa3LS0tVZyS0SCQBAAAAAAAgL4QSAIAAAAAADMtyzLleS7v/Z630Ww21Wq1Bpiq8XSg6gQAAAAAAABUKcsySZJzbk+fT5JEaZoOMkljix5JAAAAAAAA6AuBJAAAAAAAAPSFQBIAAAAAACPS6XSqTgKwLwSSAAAAAAAYgSRJlOd51clAnwYxAfc0IpAEAAAAAACwzX4n4J5WBJIAAAAAAAAGYBaGLhJIAgAAAAAAGIBZGLpIIAkAAAAAAAB9IZAEAAAAAACAvhBIAgAAAAAAQF8IJAEAAAAAAKAvBJIAAAAAAADQFwJJAAAAAAAA6AuBJAAAAAAAAPSFQBIAAAAAAMAE63Q6I9sXgSQAAAAAAIAJluf5yPZFIAkAAAAAAAB9IZAEAAAAAACAvhBIAgAAAAAAQF8IJAEAAAAAAKAvBJIAAAAAAADQFwJJAAAAAAAA6AuBJAAAAAAAAPSFQBIAAAAAAAD6QiAJAAAAAAAAfSGQBAAAAAAAgL4QSAIAAAAAAEBfCCQBAAAAAADsU5IkVSdhJAgkAQAAAAAAoC8EkgAAAAAAANAXAkkAAAAAAADoC4EkAAAAAAAA9IVAEgAAAAAAAPpCIAkAAAAAAAB9IZAEAAAAAACAvhBIAgAAAAAAQF8IJGFmZdklSZL33YpTAgAAAADAZCCQhJmVZZclnZJzF6pOCgAAAAAAE4FAEgAAmAqdTkfe+6qTAQAAMNUIJAEAgImXJInyPFe3y3BlAAAwOzqdzsj3SSAJAAAAAABgAuV5PvJ9EkgCAAAAAABAXwgkAQAAAAAAoC8EkgAAAAAAANAXAkkAAAAAAADoC4EkAAAAAAAA9IVAEgAAAAAAAPpCIAkAAAAAAAB9IZAEAAAAAACAvhBIAgAAAAAAQF8IJAEAAAAAAKAvBJIAAAAAAADQFwJJAAAAAAAA6AuBJAAAAAAAAPSFQBIAAAAAAMCY63Q6VSdBEoEkAAAAAACAsZfnedVJkEQgCQAAAAAAAH0ikAQAAAAAAIC+DDWQ5Jx7o3PuC865B5xzv7TD3691zv3H+PdPOOdeOsz0AJhsnU5nbLpzAgAAAMAsGlogyTmXSHq/pDdJeqWkdzrnXrntbT8n6Qnv/TdJ+teS/sWw0gNg8hFEAgAAAIBqDbNH0h2SHvDeP+i9zyR9QNLbtr3nbZL+Q/z59yV9v3PODTFNAAAAAAAA2KNhBpJeJOlc6feH42s7vsd7/7SkVNLzhpgmoC95h54vAIDBGZflencyzmkbJ+TTYJCPQLXGvQ6Oe/oQOO/9cDbs3NslvdF7//Px95+U9B3e+/eU3nNvfM/D8fez8T2PbdvWMUnH4q+3S/rCUBINAAAAAAAwm27x3t94tTcdGGICvirpxaXfb46v7fSeh51zByTVJT2+fUPe+xVJK0NKJwAAAAAAAPowzKFtn5R0m3PuZc65mqR3SPrgtvd8UNJPx5/fLunP/bC6SAEAAAAAAGBfhtYjyXv/tHPuPZI+IimR9Nve+886594r6W7v/Qcl/Zak/8M594CkbygEmwAAAAAAADCGhjZHEgAAAAAAAKbLMIe2AQAAAAAAYIoQSAIAAAAAAEBfCCQBAAAAAACgL0ObbBs7c879pCQn6XpJm97736k2RcDuOed+RtJN8de29/4/VJicSjnnftB7/5Gq04Hp4Jz7GwrX5h+Q9KD3/v8ewDZ/U9IfSPqI9z7f7/YmjXPunZJulvRvJL3ee/8nFScJ6HHO/bKkJyU9R9JF7/2v7/Cen5d0R/z1D7z3f7jDe34o/vgdkp7w3v/GkJI8dM65H40/frekv/De/26V6QEAPNNEBJKcc7dIuiTpZyRdlvQySV+R9AlJNUm5pE9JWlK40M5J+ktJC5JerXCB/pqkT0t6U3zv8yTdL+lFkjqSNiX9sMIqc6+V1Irb/WFJ/59CXn1z3MY3S/orSW+J6fkzSee99w87535Y0gVJr5L0Hkm/L+lP47bukfSjkp6S9JOS/ifn3GsleYXA0k2S/kLS90u6Nh6Lj5/7z5LeLOkxSV+S9ERM98skHZb0dUnfIulz8bOXY958XuHm5DpJN8S0LUu6RdJDMR/vkPQnMV3Pi++vxW0+FY/178U8eFVMYzvmn5f0SkkPxDz8oqTXxG1+Nh7T5+NnPyXp+ZIe894/5Zz7Fkm3S3pQ0q/EvP9/4j5fLun18Tz8V/Fcf188/r+Q9IOSXizpd+K5/UFJf0vS6fh7Fs/V/yDpryXNx+N4eUz75bi9y5LOKZSZryqUs+dK+nZJL5H0z+PxPRGP5zckPRqP9yWS/rWkX5V0o6R/673/sHPu5fF83xi3tRjz8yFJhxRWKFTMo01Jf0PSZyQ9LKkZz8HLY95txHP6Iklf9t7fGx80H5f0s/E4PxaPY0PSW+M2Py/ptrj/w5LW4nYelfSCuP9zkt4eX8/iv7OSDirciD4U8/jJmHc/EPPnk/Gc/L8K5ek9zrmWQp1rxP3/i5hHvxfP2XfH97YUyt6hmO5vlbQe8/GhmP5zCvXjVTFt/yCm+09iXr5Eoa4/qVA3v1PSfXEfczEvb4zn7HOS6nHbNynUgTtiWjYU6s/Dkk5J+jmFMvNahWBvGrd1StL3xvffHPPkxyX9Ucz3OxXKriTJOedi2h+I+TSnUDYPSDqjUD+uidtpSvqgpL8b0/ReSd8T8+n1CmXl7vj+uqSfiMf05XhObovvbcW8e0P8+U3xPD0U8+3Nks7H8/F6hXJxncI5f7WkNe/9Hzrnblco75cl3RvT/fn42tclfZPCKpzPU6inTYX28Gzclo+/J5I+LumlCuXvyzH9Lm7nOyR1Fer7OYXy8AWFtu9i/JyLefi1eI69QvuSxs+24/s2JH1X/NurYp78J4U695hCXb5JoYzkCuXzFoV2rhOP6XGFevHHCuXmeoVz/02S/nbcznPj8d4n6Sbv/X9xzr1Rocx4SbfGc/1ETOPzJL0wnvfnx/N1bUznK+PxPRQ/+53xHJ+V9I8V6tmnJf2Ec64bt/1wTEcrvi+J5/DOmPZ/o9DOmefE936XQjn6bkn/yDn3ZHzvA5J+JObZ5Zj/jyvUr4vxPL08puNxhbLcVajnn5V0JO7nzfE8fEGhTvxp6W/rkv6JQtlOJb1CoSx8JB77/xi39QmFenc5nqNvlvS0pG947/8v59yywvn+7yR9SKEe/oRCu7Aez9VXJP23Ma//SOE62Y55/lZJ/7ukE5K+NebBGYU27dsUrkEXYp5eVqgzB+LfNmJanhP//6pCGduMx3ROoR17VKG+PK1QDl4az9PHFOrYnyvUga5CeTys0B7dGI/3UszDBxSuCc+Jn3lDPL4H47avj+fD2qSawnl+UuFa99y43edK+pT3PnXO3ahQvq7x3n/ZOTcfz7OL27hRoSxdUKiDczF9r1Eor5+Jaf/deOwPqWjLH4x5tRE/84K4vU/EfH9Q0s/HfHujQnt2X3z95TH/Ho5/m1e41/lk/MxmPAfPV2gnX6NQXz4cj/VHYr6/RuG+4FXx9wtxf/cqtC2XFdrVV0n6NUk/JOkRhTKXxc8ej/n/hfjZXKEMLSm099fG9ByS9NGY1nmFsvIt8dxciOdjKebhV+L2Xuu9/7gkOederNDuvF3hHlUK16EnJf2ypOPOuRco1EFJ+ptxny+R9C6Fe7Hvd879VUzjzfFcZAr3lxcV2qAfcM59NqbpYwr3Wuck/dt4HG9UqHvH4j5eG8/z3XG/36lwfX17zJ/rFe4dXx1/f4FCmfmEwrVsSeFa8bRCObhVoa24rFBPPhrf86L4vucqlGOn0C60FMp9Leb/2xTu556W9KBz7s0KZfF1CuX7skKZ/qxC2bT7qovxfW+LP/9nhfr1sEL70I7peoNCW3Uonqs7VNQ1u94+rHDf2lRoy7+mEOi77Jyz+7p3xPOXxbT9ucKK1D+m0MZ8r8J9y7pCGVxVqGM3SHp/zNMvxG3foXDNWYn53VJo3z+lcN0+GPf3OYV2+UFJvxjT+e8V2vofVKin7fi+mkL9XIh5fZNCWX6BQj18gULb84Z4Pu5TeHY5G7efxvP5Bwr1+mUKZeSN8Ziui8f4x977M865V8dzu6FQzu0Z4dvjef2Pkn5aoW59UqE8Z/Fc/K2YDy9TaGO+oVCWXxO3afcavxOP6asK9eIlCvX+jEI5PxS3fUfMx7+OeX63imtZHtP4QEzj34n521WoM7+l4p7qsfj6okLbc5Ok9Xg//nyFMvpmhWe1LynUkadUPDfeENNgzzP/TUz7dQpl798plLl3xd83Yn4cVLhGtuK5vDkew3mF9uK+mBbL4+fHz6wrtG+vU/FM+AKFOnEg7uvjCmXiKwrX5G9VKCuJpD+M2/xRhbLxcYXy/1fx2LoK16dLMS0Nhfp9Q8zPhxTa2M2Yvj9TKENfUriveH08RzcptAkPxe28RaE+vz4ef1uhrfkOhbq8Hs/TcxTK/KZCu1WP76vHY/pyzGd7FnqRQhn643h8/7XCvedLYz7+qUI9/BEVz25/FI/p+pgvn1NoExfiufvT+L7bFdqin4l58+8U6saDkt4n6acUysIHVTy7+LjtJ+N7pVDWX61Qhu+XdDTuI4v77Chcb785fu6LMT9fFbefx/P6mKRfiPv/fDyv36biXuRUfO2JuI+fUmgv/n1M32fifuYlfcZ7v6mrmIhV25xz/71CY/r98aXrFBrES/F/u7l6WqGSXI6vK77HKuvF+PqB+K+r4qZvrvQZH/+WKBTE6xUalrn4mov/P6FQcPPS9lx832bcr71m234yvn4xfjZTEfhJSvu24zsYtz8X/+bi/4qfPaDQMD4vvm5/q5WOZS5uU/HzF+O+7KY5ie+37Vm+SKFgKv7ejf/sPRbEs3RbfndVPNC40v+Wlo3496+ruOBct+0zPuZhLX7mQEz3tfHvF2LePB3T8JhCpb4mfm4+psPyzvLP0pmUtudjPhyI26uV8ivZdowu/p6Xjr983i9ty0/7XBK3Z+Upi++5GD9/bdy3lRXLXzsPKv2to9CQZirOj1ScZ8s/qw9eRSNTPgbLg/n4u9UFH/dtZdHK84WYdqs/FxXO21PxdauLVm/yeI7K5da2Z+XJbnKy0nE/onBBtPJgf78mpsHOeS3ut3wsPqbJbjIt3+z8WnnMS/8SFeVyQcW5ailcPK0tuKTiQfdphZsKC9jaTclcKS8OxM9cULhodhTqvJW3RxVuDKSiHSifs/MK9cOO3f7PVbQ9Pr7fzpHViYsxHZfjdu0c5aV8tjx8QuEG4Kn4/oW4L2t77PjL7YKlNYu/W3k5EI/zoIqh0/a5ywrntFxOXdzvQvx8Fv9u6e7GY7bzZ71prExciOfBylW5rLdj/tlx2jm+XkXd75bSbefA2qmnFdrrG7S1nlh7Ym1QXvpsplD+LGBl7VWuol650n7tvNnnD8btXqdwQ32TiqDmc1S0J+UykKl4WH5Kof7cpq11NI/HYnl4MObRpZimRxRu9JJ43NeqaMusLbR22H63sn+p9Dlrdx+P6b89pt/a1Dym6UBpe+U2ybZpbYkFN+y9dm0vl/ny9dza4XI599vOlbXHth3LIyuXdiyZth5Xr8dI3NblmA/fiHlws4prjF0r7bjmVVx7yuf9uvi/tel2LBfi/+XrzkUV5di2b9fw+ZgGqbg/sXqg+LdrSumx+m35Wb5u2c/W5lh9tDyyNtOuYXatsfy6TluvMbYNY3+ztsz+Vm7frH5aXSuXNcW/aVseX6NwT7Gook2369FlFfdvl1Xcf1l7a/u0tqt8D2BpvkZFe2R/21BxjbPXbR92vbF8snstu6/bLB3vhZie8r2rtWd2jHZdtfJr9druXTtxe3adt3wrt6eutL3yvcL2ey0rLwdU3E/ZNdvKp5UtOxe+9LqVeXv9kor7ZMuny6Wf7RxYW6z4d6t7XkW7cVFFUOTpbflePmd2PbZ22cq7XbvsQX9ORbtr5dnSYm22pT1XeIB9ocL1ebGUf/accH3pOOz8WXskhfNm9zAWmLT7n6dUXNfnFK5hi/EYy/erdo69wnm366e1a3YvaWXP2la7P7B2zc7h9npk1+trS/sp58f2e0MLwFi6y89fdj7tOtxV0cZIRf1PFK4bds22umb3S9eWXpeKOmaftTy2elO+976s4nnD6nK5fS/fG9q285jO6+O+7Rpi+Wpl0c6zlSVj6Xkifr58jbF21H62+p3GfTylcN9hX849oXB9tjyWivNkbWy5vSm3DVZvLZ+sPbB8sGvPRulzdr2xtrxcrsrtdfm5pXwNtuCDfd72eU1pu91t/+ZUtB32GTsnF1Vcbw9sOz4r/9aWWTne/uxdrvtS8YxmZSFXcd9zXSmfrWzZMdhrVm4s7VbXDigEsm5S0bY/X0Ub9UjczyEV1wKrTxsKZdTu11VKR/nZvVzerM4+FfPdruN2zbH/r9HWemJ5VFNxz3kpHtdphYDVh733f19XMSlzJD2mEKG2bzXtInROxUXiMRUXkksKD2kbKoI0j6q4UCYKUVg7GRcVMi+N770U/1mhtBsAq7RWOO1B/xvxPR0VhdQq7iWFSOCjMZ12AbuscNIfi++xb3IuqbhZsEbnyyqCZnZRskJZviE/UHrPJYXI8OMKjdCXSvu/UNq+3QRuKEQwL6t4OLSG46yKxtIuvmYjpidVUSmtgbMbi8sKF93z8b1fj/t+bnzPfPzspXgu8rifa1RUULuw2U2JVeCvx59vKh2/XTgsWq7SublQSofdgJQbggvx/ZsqggB2jqxMfFjFObR0bsbtPKWix1M5Smt5/GT8ZxHkA/H4yzfzTqFsWyP0dYWHMmuwLf12sbNvkbpx2+fj8XxdRU+lizGNdi4247+Wikba8uKSQpmxQEsW3/PcmN6n4u+PxP1ZL7mH4766Ki6wVlfsvFlDaGm3vDmv4ibzZoXG9aul8/ZlFY34cxRurhS3+5iKGxx70Lgc8+6iQt2zum0XjE4pbXYRtiCS3dTdqFC3n4rb/ZJCe7IQ8/bxUn7aw7l9o2PtlFcRbLZtP66iV88DMb21uK+NuG17KLgYj+9rpfTbhcQuFu34+qMqHhws6PGEtgZI0ngsGyramadVtG2L2noz9rWYFisf1mba8bi4rwOlz10f/z+rIvB3QUU5L99UW28P+xbQbi7tQfZAKX+tPbL2ycrU0yrK9uMq2pDnqrg5sQu1SscqFWXAblq8Qpm+EH+360s35sP5uP2LCuXCqygHVubsvRsq2gHLH2sz0tLv51XcrHqFb6m7CuXmoopgj920f01FL1Zr9+3BJFf45vFphW/5nojHaTflh1R8CWJ1sKtw02qBCzu/dl7s+mjpPavQPtk5sPyTimCrU3jYsrJqPR4tmGflp6PwLe5jKm4+uwpl2crKN0rnxtpqy18LUm5o6zVtU0Wbl8c8s2vlIzHPrQzaNeKp+C9T6EksFQ8+1iaU6/F9MR03KnzjWQ6o2A38hoqbY/t2/7F4TI+pCBgtqgj4Wt4nKq6fUmhfLX/KDw6WBw/F7dvDQhaP7esqrkVf0da25REVwRO7f7L9eYW2xSncZH60lLYNFd+KpyoCBxfifu362S39/qSKMvqVuN3HVbTZ5YfZ8pdwdgNv9agcdLAbabtHWIjH9mTpWOzcWztk10kLVHRUPCjcoOIesfwgZfdmdvyPxDRb+5GruJYqngtru61M/XXMK7tXuF/FfaDdr9yrIgDSKh3j5Zjmp1UEga1HnAUk7Dp/WeHaaeXwbhUPRp9X0UvEHqDuVXHttc8cUGgjym1mJ/5uDysuHp+1PU+oCE7bFyuPxuP/cz3zemX3O3YMKuWNpcPKsV3H7WHI7sWs/FkZs22nKq5L9kWdVFyP7Jpo9dmCBnMKZdLSYfdJX1Nxb3lj3G5dxXXKei1YmlRKt305aXXugkL5sWN5QsW95mLcn7UD16kIANq1zL5YsXw6GPdn1xUrb9+IefZE3M8DKq4nT8XjtF5c9kXhJRX19CkVZd/amCdVPKM8pVAm7EuYjkJ5eEKhrUpV9Ay0+xIL2EpFwM7uZe1ewu6L7ZrydRXXhsdUPBdZXmxo6xewFnyx4KPd7zxHxT1MN6bTAo12T2iBkodV3JtbW2HXGysrdiwbMW0WiLb7atvec7X1fsPuQb2K+wvL82tUBBmsrNo2pKL9Udy+fe5JFdc+xfd8TaGNfqyUz+UvOawcWVm3ulO+B7Ofn1Rxv2j3Q3a9sPZiU6EtsfbY7i3sfuxeFeXN8q78xVG5PFib95iK56Y5FcE4a7vt+dGuNZcVrrOPx38X4j6t7khbv6Cx64H93TqbWHzhGwrl+LyKc2/H/URMn7VVFrDpKvQ6suuwBYyeUNEeXKfiuutUPBMvxvdaPbBnAbtGWT3pqKgHT8X/Pxb3Zfd4tn/Fc2Nl6X4V99BScV0sf0H/qviZR9SHSemRdIPCyfwpFYX9JoUHu+9V6IK2odCVbkPhIt6O/x9V6Mn0RYVu1B9VyLQfUrgRvEVhGMi6wrCS2+O22wrdYDNJ/5tCI/QmhYLzAoXuay9U6IpnXfTuVXEzel6hW+RvKHTRvkZheNGX4v/PU+iyeCpu44JCN7iGQuN/U0yvRTDbCt372goXsD9W6AZ6UOEifVt8/euSXqpQUH9DoSv/xxUKyO0KXROXFArrX8b3PqrQDfCh+LeHFYYmPDf++2L87JxCZXhZTMcfKAxBOqfQZXIz5kktnpPfU4hq3qvQ/bCj0C34t5xzb1HoEn085s0PKHTp+2rMjw/ENL86pu8xhQb2EYUuqz8Y0/7rCjd/r1QR1HAx787H/L5FoaI9rDA04rRCt2sXz909KoIFBxUq0bXxPV2FaPI/Vygzr5b0ryT9zwo+pWJ4VDOm4aGYjr8X9/EnCt0q/zIeyz9SCIw8qOKC9V3UcUwbAAAgAElEQVQxzQ/H97fiZ26I2zqrcAOTxPS/SkVX9a8qlMlvisd8g0JX865CV8WaQhl4tUKQ5kaFbp33xm0ciO//VYUydDB+7jvivr8Y9/H6uL+Pxp/rMe9eqFB+XqpwI7+q0EX34wrdOV+vUG5eFd//6pjezyg8gB1SaEBfrOJCdqtCV9UfU6gvfxmPy4bT/VjMQ7sIPKVQBj6q0Ai/WKFeXBvzcUmhzH5Zoavy/6JQNt+i0NV6UaGL/6X42v+q0DB/p0LjfYtC+f5zhbpnPYleFvP99phHtyhcwP4kntOH4vG8U+FG7oUxbb8Qj/E/KdSxX1MYRvAVhW7lazHt7Xg+OjE/3qxQTr6oUG/ujPn5dRXfMH5Q4Zx/RaGMdRXaCrtwLSl0235ExZC7NP57Xcyjr8Zjem1Mn11Yz8X9PaTiG75MoV25qFBXajFdj8W8v1uhff2ZmHePxM//uKT/El//JoXu/W+J+XRj3N4vKlzwXx+P+w8V2vyz8VzfFPPzPhUP5PPxOL6oYlhvHvPy9pjuaxTa4I/Fc2YXY6dQh2+Jx/VEfN/zVQyJ+B6Fi7EUuuificf1VhXtwVmFOvyWeN5cPMefUBgG81kVwZNUoUv1fDxfFhhdjMebK5SlTyh0Ef8nMY33xuN4OKbveSp6ot0at/XHkv5pTO+6Qll8eXx9I56n8jfdj8XXPhE/89vxfFg78SKFa9ZvKdT/71QYanZExUPMQYVhDPfFfFiO+zuvMHTlokL9uE3h2viwQpk4o6I34iviebstpuFahWvj0wrtxvepGDb9TxXq3Jfica8rdOU+KOndCt24LfD2kEJZeySm7ZdjHl8f9/l+hXb7V+P5bCjUm4diPtwU8+XHFdoy6xG1EPd3IJ6XVyiUn7sVyve9CnX6ZxXK+U0K7eU3K1wTfi/mxR8olLNcoV38FoUylyjUvVfEc/EJhS7xn4zbe4dCHXxhzGP75v2fxNc+F9OxoVCmH4rb/zsK5S6JaTyjopfUy+JxfXv8fzPubyGe5/9T4bz/qcL1/9H4nhfF8/CUQvvxaYVhlXfH/dyucG/yWoU691YVQ+WaMV9zhXbqfHxfEo/vFhXDA20Y5v0K7cPJmJ82rOQv476+R6Eu/O14Dj+sUK4s6FlTKFP2zfzbFMqDBTo/qtCevEKhvLxXocz/tEIZskDVxxTq/z9UEXj9zZhf36dQb14S8+HnY/r+//bOPNqvqsrzn/MyTy/zSMIQCQiBIIMWIKWxVYTSgkZESqtKxapq7bYpV/XSpXZpEVSq0SqWCtioZYk4IXbRAmorIhCQMBoCMSGBQOZ5enkZX4aX03/s/c0+v5sXCFq20pzvWlnvl/fu795z9tnT2cO57/UxrPOxP0A4/6ucnq/G5P5pv+/fYLwwxse4DbgZsw3vwHyvk4Cb/Hfr/HrZoOF+/01Yu+cKH187xisDnR5nYFUQr8VkWX7DGzAb8T1fu6V+3Z1Oyyd9nu8HrsXa5m7DePIvMX6RP6CW+JMJ3nib01RB7UWYTXuVj+tUTIfdhPHqUf69qf6dVRjPTfXv/wTzN97oc1uP8dlQp9UJRFveRZhvsAeTqXOJgHt/rC1QSdChmDy812k4FpPLezEZPBLjsfMwPn7Q1242xr+D/e9H+xivA670uazF5FJr8V3M/1Rb/lO+lv0xX7gX5j98COOHB3xuYwjfdyCmf7Zia6/A3GJf95FOg/2Yf3uGr9NPsVboBPxnTJd+1em+huChzxKdENsxf0xtpuP9+gWY/T0Tk525mHz/wum9ArMPp2E24xxML53jv3s/1vb3PV+XZ/y+R/nY8XHf4ddP9XsO9f/PxHjjWCIxcynG++P8um1Oi6FEwnGhj+sop/vDmAw8humX1/pzHvd1bcf49GuYLjnZ5/pBzE5OwHzJv8Z461aCV//S11WFAK/H+E37lqv9mQlb5/f481YRAYnJGM+839d3vtNCbcfDgI9i/LzI6XKyP+c5v9cOot14vP9+A9GuOx3jhTGEzya+VaXmfkwHLcH47nuYbb4U450VmGwMdbpsx+zSOH/OVEyvnEXYhr8gCizKaswtwEcw+fmBP+cYTLed6zTajOmvuwnbdpKv7cO+Zr/23z/nazSKKMSYgvHGLF+HfdheYKhf8zDwT5ic34DJ9klEW9xq/+4lOed/Tim9xq/9DOYDTsNs9p/4d6/H9grHEAm2gT628T6ehPH2f3daDcDW/yJfwwE+5m/73Pf6mJQQ1Ll738ViHDOcrl3Atpzzdl4AL5VA0vUYc03HBEClYMri9CfKthU1hMgEyDArGtyLyAIr8zSEiISqVUwMWmbbobX1QRmyZzGGUzZeUU09bxARnVUmQhFAKauBROaozArvICoFlJlXhLG/j0kbgTKTo+zICCJSX2ZnVH1SZu3L8ki15KisXdHUwUTGR9FoRfKV5VU2VWuiAJzakxQV1vVl2alaGMrqELWhKEKuTZ/akSDWT2XZZWliWRkDwT8yfJqz6KN7bSSqBVQyqoooORcQWVy1oChq3h9TXmpBUbm3AiDKnCg7O5zgDZUcam6q2JCDoeyCMrqDCnqUGSCVHvclNj5girwNU4TKNqoMWu1hauOQgldWagCRNVZpqUq2IXghEXywl2ivamaV8XmsI1oUO/1zWZ2nqLqyE2XLqVoF+znNexPnH5R0Ev+rUkEVHAomji7WRllaZQnEP520VhpBZMBUcaU2H31vqdO6lNFEK68qS6tKSVXr6Rq1GUiulCXagfGdyvdVEaZ/4gtV0Cn7pSxNLyIrOYTIlqnqT7KpyhJtmrUmyuwMK+4HUTIrPlcmSG1Wmp/0RbNsV+XI4hFl6ocT66gMUSnDqujQGJX1UiBe9kJ8Kf2ucSmDXFYqgRnyIRgfSw+Iv9VKqCy6Sp2lY6C12qOXj0eQA6pzbzS2snUGYm017j1EO6X4osyqjSH4dwBm+/oTmWQ5fJJ93Ufy0pto6yozlqo81VpJVyvbJR2uEm2NTRWx0qXS33swh11BK+lLVfpoLGW7hHSnMpmSjb3Y5mkcoRcSxv8LMR44BltLtZhJT1D83I3xmfSV5qAWhD6N6zQOlZr3Lr6ve5YVENKhZcWe1k8yud9pKNnv9Huq3UK8Cq2tPtuwTdgJRLuibFO330eVNiOIIwLEK1qTfpgOKHW8xo+PdXAxFq2RxiG69XKaKFsquyJ9BJFdLm1HXyLYpcoE2XnNXfSTvZJtUcWiKnlKf5Difgoiyc41W9GUFZ5E6CTptlIG5M+Ulc3ttOp46RPxruyFvjeQqHgo2546/TtD/FrpJwgZ0ZlgOs+urAoYRdgQVdI0K0xlR8rWI7WMK/tfHhGh+e4vPuv30nGqApE+7yIq6tqJSgSwDezIYu7bsWDGEKIKaBetvtOAYk4QFTgDiCqhzMEtapJP8b1oWVatiM+ks2Tfxxb3LW2TKhL6Ft/fjNnFtdhGcrz/bSxRoTGquF52S2uqdsGyTQhMvtWJoXN6tI+R7GgtypYetdirMkg6Vnsardkmom1cNJOOlXxB2KFE6LOu4r6yg822QukotaKWPm2v4hlaM1WvlTpbfnIq5iPaq5pK8rUV4y1o1cfQWhGn75f3lR90FLF3kL6XnlCboqonZbM1LrWsya/aTuzNRF+NRQkhVdnIn9C5XCXtZUdFu2a1k/Y5svPaD2mPo8pL2fbyyIpM7CfEl7IxWhutRbmv3kEcV6IxlX6KdGbZybOH8MU1P/kA+ine7vBrpNOke+X/q4qsP1Gxp2MKdhffkZzJ35EOK1s6xcsaqyrBZeMo6CafWjpKe2vxutpeRT/tvyGqOcVLop3uL1nqU9x3HxZ7OLq4RutTyqz4XeOWLoTgMfnEGdPDC4C7cs7X8gJoe6EL/kCwC4vKludNiDkHEmXNKqeTwyjlo43gMkzxgilSEVcCru+IwFpwKRQJc9kW04UJzyuK8elaZZa3YUp9O9HCtgXLTC0jot8Qgr8H2wQkwoHfR7QcSUi1SVHZ5X6/9zaf8wZaHXxtfCRI5cZAG1MJvYy3MjKqbujGNsVyAOXYb/a/64C8TkKwxhAOrDKBG4nSSjmgjxfz6CQ2DF1+X9FBrQ/9iZJvKVAZArUc6vsKgqgNQu1p5eZHSkwltAoWlIK5C4uWK/DQ6fdUMKN0DrV+MjhSvirpl6Pbq/he9r/LyeuN8ehowmGAyGS2Ec6UnI31Psa9Pj5dv5poFRmJKVbxk4y3AjX7iY2Hzu/YSARMH8V4Xkqpm+D51dhGbieWjdBmCqIUdxnh2N7rcxjk35PxK9tOxZ8qw9emvau49wN+T1VWSWb1fZ31oACxzoUR7dUiqTUSvw3yMcspWk1r6fR2v34IoVNkyHUGSv/ivhq3jFgHrY4YtBrUshwfTK4hAiuji3FsLp5ftqQpQCknRe0f24nWuGG0tnE0AyPiPRnQ9QR/qjVwqX9feldjkDOnYFU5V22gtxKbJwVhZNC1qR5JnPWww2mn8vZ+hNNdnseg8nw5YgoSU9BgGyYPcjwS5vwrq6qgjlqaFUCCCGCprF7nePQlWrmepnXD1Y05TuOJdk7JnkrFOwmHZFmxDmohEt9rw9ONyZvmOx5ba9k76XutiZwizW8PJn9y9OR0yUkt26+g1dao1F4Bq3WEs7+R1paGMuGyG5OnPVgmTfSVQy2dW25Cd/o/nQsg3ugi2p1VHbsNy7gqUDMNy5ZrXo8QrXCyT2sJfSMeeoiwK6Kb+Fd6SusBEezZQ5Tn7/Lvd9Nqk+RoLvHvykcQH++jteVdDrB4V218m3z+kr+ziXbcIURr5jwiIXGE/1SgQg7uLozntClYVoy3dNCHYDKijcY2WlvQ9mHZT23GdJ020aJT2bIg2ZK8bSMqBCUP8mvUGgRhtx4hNuU6d6aNaBWQjtPGY7g/RxtBybM2UO2YjpWehGhxhQhe7iB09SAiISMZkB+4F9OTail4BOO5Tkxuyo31TB/bKGwdexf3XEXwwUof4xjiRR2J2Iht92c+QMhdqa92Ev7f/QQ/LSIq7ZZj2XVtumQfpMvla8jmatMtP0xt92VwrUzWTiICeW1OizGEvtCmSBsf6a4dhB8ylKiMVVuPdIbWaKmPWf4jRMvm/IK+mpsOxT+a8Bn2E60fq3z9lFDZ788B4/NemN06BatUGuPr0odIvogG24u1EX2VPBZd9/i4x/m91IbYRfjpoofa0eSjSffLL5BPolY1Hbmx2q+ZSWxQy3ZVinEqmKu9gnhyE2ErlhJ6T8FhVcCp3VLtcN3YBll7k7U+j36EHtTz1xIBYfHGEkIXiI6jiE26eFJ8obl93ceiuazHqo4GEsFA8az8A/Gixqp1lE8ziOB1/ZTO68R0ktZN9mdQcT8FL8tA2w6f3xaihRLCD1AQVj6IggRqx95B67EN3RgPlUeRlK1aNxH2UX6w+FsJXdlC7Ue0nxCvdNPabtlFnM85jHgL9ArCh+tX0EH8oYT0KEJHKFnQ28e13r+/DVsjCHu+nvAJN/nv1zhtpGMexvSpfFIFibSGsj8ri3vIniuJLD7uxuyfvqfnax8u274f07u9Cb+prbiHXgIjGdLzpziNFvs85N9pfy1dtYiwWXswfa49qeRiNeandmHVYOdzGOg1Y8aMw7nu94orr7zyEazkW5v29dhk52FKbjmtB8ptx4zlnRjB7sTKjU8gzjB5FBPMFcSJ8eoRnoUZiz3EwbNSdArcKEuijVw7xiwLCKdBUcqVhJOokkqVm40inBYFNOTI/di/fzfm7K3FyrvbCcOszIyCGFJijxMVJ3dgRrCMui7GBHiw020WUdK+E2O65USgYYU/4xkimzMKE7I7fe5P+RgX+fzn+Vgfw0ooj8XOnngYKxPvRxwwucKvPdqfq6qBTATfRhFGWWdcyIFUxU4frHxyJRH80CG1UsDP+b16E4qgy+ey0Odzn19/OxEQ0vlGy7CSR2V9xmH8sBwzjDJAykC1+X3XEG/CUzAmY3wx2tdWGaUye7iCMDiKbksRrfS1noXxiTJxko0urKTx+5iTlvw+s30NFIDcQgRM+juNFhLn9vT2a7ZjJdTf9mcegynyYwijrlLXKUR1iDY4ynYqK6Zg1UAi8/NssaZricPJ5UQtcfqOdFo/jDlyOrtgJJH50cHuKzHZXOs0VoBTsj3b77OWiNjL8GzDgr63YOWfY7EyY5WVin43+7rNKp5RZlgVhFmFyfMmTI+1EwG7DqfnFr/HAEyWlhDtXGMx/pyMleFegrWG7neabiJKsfsRb+JTZZAcCzl/WzFDvhHjQQVDB/haaFOr7J8cMOmLHZg+Ub+3NiTznW4qbVfAZ7vTYD6WkVPp+Q6Ml0djZeAyvAp+iaaPEuc9SEfcSgRZ1c++D5PJDUSp9Bz//CytWZmNWLvVw5ge68J0kc7g6MBkfg4m0wsIuX+AqLZUsE0OnlrrtOkRrWV3MtZ+vNnXVdWnuzF++gnxhs3RWHm3dNdyn6daM4f4XCdgOlgBb20+Fci8x+93FHHgpAIeEJtlbRBWYGXaJ2EtCFOwTcFijI+1Ierl9NiE6ciHfBw6D6AfUfEyv1jXFZjcfYwIVu/1+fZzWs8n+Es0Xe73XefXDPN/XUQZvpzdYYQOuB6zRcsx2TyCCMAv8/vv9GdLDrf4eHZgLRez/L5y+p+k9cUCspH9iI3JYB/bIoxnhhPn+WijqYpfbTTWEAEwZSxX0nqgv563BpOfcYTDuR+z2f+G8bOSXeN9fVYSm+EJRFJkNcYbI4mD30cSwURoDewrMLAd02lDiOSP/B1V2MrZ/4ZfewJx5sMOv6bDxz2WsAt7sTab5LTTxmEzkYwRH6/w62/w7+x0mujfOkxH6oUJczDZWux0UlJMvK2q4vswXpxABIVmE+1T4lHprQeL6zVvJek6nK59MZkc5/c/vXj+Xv880ddmM/EGnw3+e200JL99fP2O8v+vIWztMGIjMxPTg9KLa4gzJI/3vym7j89pnP9b6fe8hmhlkW3YgumATiKBuMFpe4ePbYQ/b5DPRxvLbuKtq+t9zM8V81Jipq+v3yx/3gQiyKHquy6iokYBDh1mKznQGTLaoKv6HMK37YfJVR9aqxuUpFWltwKNbcQZYUoObCX8GOk32QQFDb7g3/0FEaxbReg7nXU6H/O3HnWaDiMqRlXhqs3iVkyOdHxAm6/LKp/vYox3e2HtavswHvsXrM3uj/yfglk6s0YJiYTpvrv9uQsJu5P8WlUh7vc5bPFnbMb45MtYq9M4TDd+EWuH3Ojj7HA6thNBoOWY7rgLC86pClEJCIhgmnhGR12swPSHAiLy18f53+/1+zxMtDDuJXScApBKCC32+yzB9KoSaJqv6N6B8cwyIpAhZMwOPYvxw1qiynOZ3/tup8lThGyrMnwn5pcPdLroXDzZMSU5v0K0vMsOKKC/lNjDZB/LXkz+bnVarCLeNvZNTG/8GuOdc/3eOn5GNH/Q76czVnXm0CiismgBEWjsi+nix/x5txGFFgqULcZkU/w+jgiIbsH8dMnZUsLXVQfDIsy27Mb2MnuJfcd4ooX+Gz6PZzA7pWrRR33uP/O5347x8JOYztPZSpsxuznM/9aPsAcDiXPlFmE8+LcY7ytRpEIUFc3MwfTMU4Rv/w2sLfoprDVvodP/F5gfvRk7VmEKkaAZRewv5MvP97l2YHp6GfD5GTNmKHFySLxUWtvehRHqLzCij8KIqAy6yt7aMCM1DiPeTuJ1iMpc9iPKw7NfdwvWq7obyxaoTLrMHrdhxnMSkQmWEVEWdgHWx60o/0ailE6MLMdF1SbP+dxkePZjiu54TKin0KpwVBqrNzppU9fm89TJ+72IM1GkUJdhykolsHJKR2PKSxFsZRKXEa9ZVwVDl9N3FabMdNBcb+zcgLf4dWOJMthy46lMgEoHVVIt5asSbGUd92IMfhphwLQJlgOiQ/Tk7ENkdlSVpko1ba6V3SnfyKQM834sUHA2EelXeafurc3lZizDvZ3oR1fUWuuiTJfWYgDGjwoaKcv2HLbeXdg6H0lULXUTh/aJT9SypjnuIjYXUpojMGW72ueggCJEmbwqYTb438sSZNFsI/FKY2Ui12G80IYp8FF+nc4ROZLIBqvcXi2aMjQQQSVVP+C0HU+c1aNM/EZMMasVSpn0NsJRk/OrzI8ce1XQiYaquNjt41KlgQKU/fz5EwnFrkBeWTGx0+kmB1cbuL6YQzAJMxRjfJ7LfW67CCdFVW+r/TpVXClYs87pqcCx5rEXCxqeiDkQRxO6IPnvXkFUAS0kzgfbRARDdL2qvRQ8VOZNDlFf/94kooJwKpFZSbS++UbBAm1kSr2p6o29BM9qk7HHnzOFCLKUFVoK9Gm9y8ytDPAcTB930tqaNxBzzCYRlSj9CL2pDYCC2CcQZcVlVdhejD/GEoEazU1nlclhHOJzW4vpi/mYLlbwpZS7pUTQ6Vha3+Kj0ufk100mWid6Yw7WkcQrb1XdBXFWncrPl/s12vio+kU6dg/m/JyK2QlV240gZFebveOxjaFK9ddjPK7NhGRT2WdtarWxUYZ6s9NNFaMDiWDlYCJ4IH2kg4llXzVf0ahsyYV4m6ACPeJL6Qhdv7kY0/bifuLtsopLMiybpgCnNtBKCiTMsZtWjKcXkaAYQ+i0rRhPbCMOz1zrNFNLlZ6pnx3EGR9KgCmgo8qQudhZDdIfsrPSaSsxHYJ/Z7nfR4Hg8U57JXrEN7KhO4k3QMpm7sOCFqf6d67BzqASXyrQqED3LqJ9Xps3+WRlhaYCKapMkiypjUP2X3ZXCSEIXSHbq0CfqiiG+t8VkFmOydUSn5/OxMi+Lv2wzYF0wWpfP/Gr2kWks8TX0hnLiQ6BUf6Mp7Fkj5KW8o+kD2Qj9LmTyFB3Egc3q1L0PkLGjiPOvZIcqW1NFdRKOszGeFZJ0ckFzVX9C9E2s4oI+g8kMvPih7LyWll8bWgGEcHMgZjdU5ViGZjZ7N9RIlT83AvTPaVcq1phqK/LEX6N2uo6iCSEWlLb/N7ri7HLX9tE2G/xo6oktBFXYGEaJsuqjh/r43gC8/uXYEEhVQ7I3mv/0UH4QPLnFKDeQgQVlSA8hkgsi5/kJy4hfDJVQcjX0BptdfrswdZdAfeTCR0mndSP2NRPIoJdkh3J1j5CHlUl3stpUvKPdK0CKNs5ODnUTQQfZefFh+ItzVt+qnxjncOjgJSqzVS9MQo7G/O9mP0ciflHXU73uVgypS9WQfhqIhG5h3jD6m5ibyKb1V38TtWs6/07q4nX1m8gDvxXcHgVcXjzLmL/U7YuqcChrCKSzKg6uqRVSa9uorJH+2UlUBSAL3mqLHzQuir5uKu4fiPG7zoaoYtoDYPQDWX77gYigbHSxzaSSH7uxnhUiVHpdbULqqJLBQazfZ2GE0Gdbh+HkrQdfu144uB8+bJqI5f/30b4wP2wPfIo/347kfSRb6C9Z7tfU7a198Zsx5GEj6Q1bfPxthMvNxqD8cKYYr5aa/nz2oefjemCOZiu34TZsV5YfGEilkg6i9AzZYL5x5j9uifn/GleAC+ViqSLsAkrqi7lL4bciS34GlojxzJmysg/jTFfOxZ91GHSr/G/jyIYVQ5eG3aw6KmE0duDMfwTRGvHOqwS5pXYgnQQDvev/N5yHDb6WO/EoukJU0zPEgeiJWwhFUFUObn6YZVZ6oUJnJTPXMxgTPD/P4ZtHDv9mgFEK9huH5eyBgmLek/w+0ghyDFdX/yc5PP8EDCdMGyTCYWjElApdjn86hteRWxWP4o5uDsx4Rrvz5ExUZuTDKk2PQuJDJac/KXYmqtcXhmQhEVtE6bgVMEkftpDCLmqZFROP9DntYjI9OzA+EIVIzJUD/g41Yq33Z8HwT961uriOSuwtV9PlOEqMDiEiORrngpAqC2n3ceikk4I4/UqWgNdQ/waBQye8GerQkWbqk1+j+EYn8nYzCLOLFDg5Dv+N1X1jCnWTcp9EcaD/TC5mOTXzfV1GeHjXeRrO83XVxtTBU60dmpjUZuTDKCquBRQm+f3mYhtjgdj/P19TD4GYrwhB0cl+n2wyP4Gwjm9i9hw/4SQx79zOmt9ujA5l4KejfHGqU6DXxJtq9ocKks9rxjjX2GHNu/HMiDDibdSfh3TXwrSqPJCbTETiOykevqfI87IUVZ1OMbDciDXO106iIPae2HrryCuslzSdxCbYFV+qax4O7bG+zA+e4SQz3f5/dY57UdhzqnaM/CxqkrhF36PFcR5CXJuFxAbwgW+NgroqNJsrc9FstLfn73Or13uv/uZ06sfJqeqzizb/cYRLXv9sWzqKT6GVUQb01Kf53SMz5cTZ5hoQ/4YoceOIF4w0MevV9BUgWHpcm0It2D8tI/Qj09hztM6LECwD6tOGUsESPc7/STLCjzozIeZGB+pFH+Ej3Uy0fKzAcvSHUs4cYMxXhxRPKc3JhN9iDLsdURLkdZ7b/FzLcYTXU472f0NxBstv+/PBuMdnfMlZ1BJpAewDKmqg7ZjlV+v8M+3+332YvpnG+HcLsbkQhWfI3xNFmL6ahax4ZxNVAg8iq2nDlnF574JW8flTsvFRHWFkhRtRPuEsstqxRhEtAHID1hM2MgFGM9uIGyCAh3dGH9+nwhadDudljpttJlZiWU9+/r8VA281sc+kXiL2wLiEHudL6hA8CM+tvMJJ1hVKJlIWMmJ1jzURtGNBWBVRSuevQ0Lnqh6tT/mS8mp30JUHyhBJj/mTqf9TkyXnkFUKHT69T/F+P0UzIZMwHRoxpxz2TDJvHStKsPL9dQGZRdwI2bfOrCqxNf7s9dgQZ9TCfkegPHmfKfFo8T5QdqIL8N4dynmix5BbBDUTvcqv9d6v3a4X38ExhdziDeQLSpoMQTTbSdgPqKCxev9HvJxtflRZfE4TB/MI5I//Qg/sx/GQ6oen0lsGPcSb//ahx10fg4RUFfQLyWIj9MAABVoSURBVPlzt/h6/dLv+6SvzXqMz3Y4PUWvR/25Y4nWLp0dtI2QgeF+b3UGbMCqP5TVf8KfP87prAy+En8/xta5G6vyORHjWQURdvg8VxHBs8GYnTuSOONObTlK8sm3Vbtef7/PSKJ7o6//rsvvtYyoapM/qH3NEKeH5PpY4liBMcQbqtoxvh6P6Z9RtJ5zpkoL2eENGM/9iEgM6YgBVVqULXNqN5S+h0haziGS4apsXuBz68Z8jP1OV1W/K2izE5MtJYnk264lWuEuwipLFBwahx1g/07Mbu7z58uXeMLpoPMnd2M6ZaTfV218P3D6PoLx2yIfZx+/3wW+rr/EeHYVxjtziAOTlbxXMlotw0N9rGv8mdKZnb6OXYSvpqTk40QirW9xP1V0LnJ63EMcX7EIW28FtdSJcxVmA4YR7e1riVY2JW8k/wo6rfPx7yIqeWSL1B6vCsEurALqBL9mlz9DvnBfwv/X4d29fI5lcEd+n/TNVowntQcb6vNcSpwvrKTQYB/bcAzqHGonWpN13uMtvlZKim3F+F/V2McQbentRBvdFqIyaCSxT57gNFqLycdTmE79FbH2R9Da2jyAiIXIv8bnkDFdNNDXczHhQz9DYeNnzJhxLy+Al0pF0ljsTRl9iMDPsxjxjscEZh62OCOIct2R2Bs4ZhPR3g2Y0/hxjNHPxAim1oTjMWN2AfYWhXdgRrk3pmwuwhwPbYYzcF/O+bqU0lR/5p8Tmd49mMNwAbYhuwRzRtbknP/Z53d5zvk6/3wK5vCfhZ0234kx5Eqf163Y2+RGY4J1Qc75r1JK3wFuzDnfnVL6e0yhnoe9TeBI4O6c8yUppR8Smwox6F7idaTXYM7lXU7f43LO/ymldBWhpNblnH+SUro75/zGlNIULJJ/H+Yk9saM7hR/xhQfz0LirVAXYgbgyZzzlzX3nPOT/vmzTv9LsDdNLMMY/4ac89kppRmYU3RDzvnOlNJbMWH9tK+1zkgYTTiI84hXlP419kaRf8KUzVLC0XsUq6w6HVM+NzoN30Qc4vZozvljKaWv+f8XOb1/4PN/wMd+JPaWtz9zXpmC8e9Kv0ZR6HU558+klK4DvpJznp9SkkFbivHrF/y7r/S53epr9nlMYajqbISP8/v+3A9i5Zu7nRbnO30URHoNpvhG5py/kVL6FgYFx7LT70l/3rnY29OWY29nU8vYIsyYKICqarfjMeV0JuZU7QfG5Jw/lVK6EJOxT/v8dgDP5pz/3vngH4igzmlOizOdrvMxnstYue0VTt/VWOkuwH6Xzddj/PcJ4o1ux+WcP+xvEPxbTM7f5nO+3cc8gXA230hUnIzwOf8ACzhcjhnFMZiz+4iPayrRYrsw53yf67N/wQzCZGBezvk/ppT+0cc+DMsAbgN2+BjvJrJOwzHeeTrn/AWX/Q7srSnf8mdNxHTdU/6dgZhcPNTQNcP8mZ8lqiVv9Tmfh+m5H2L6aEnO+ROu587JOX81pfRfMX0GxpdHOl2ewHTCel/TyUTJ9QqMry/y+/9DzvlOvwe+VnJKTsRK60vdNo/g1TuIqsWjMAfpYuCpnPMHU0rn+drKgc4550+mlK7xx/0c0y8K9o1y3ngr1mrVCczw8av9QRnWBU6nRZiMX+Xr+GlMVsZjsvZ/Ch13IaGDvoTJw1Dgj3LO56aUPuB0/VPMZmhjvdbHsBDTqW8m2lYViN+BHY74FV+XV+Sc/y6ldFPO+b2FHT0BswFKuowngmdLfOz7MR29D9Plc7G3Xv0jFjDe6ut5H2FnH3S6fRSzL+/Eev7/2O93L8ZT+Hele9ZjQdhvFmPZjdnkddgbTS4lWmXnYLZe2cPBWBbu55gcdvoc8TWf6886F+MT/BnPALj9mIrJzeWE4/osFhx6Lcazb/a11ry6iFbwYzAeG4vx4Uy8itTfznKa32cllvF+xq8bhOmH5cB28XjO+Spfr/djdgin3xpM7ue7rfiZ2+IzsTfWtAH7cs7v9GfiYzsbf92388cUTOafw+zol5wOl/m17ZiOaQd+5X87Lef8OEBK6S0NmZ3qY5rq9xuOlejvwXh0Q/HdKzF5U4ZUSYjtwJtdPi/HAjfdmE590u93BKZv7nMadmE65T2YE/yvmL27BeOX0zBZnYTp6nl+/QWYnliO8dlzxJtJb8Uca3LOPynmeBqRfDsJs0Xn+Lo/gbXoqG3iu5i87sX4ZQ3Ghz/EgkYP5py/llL6G0wm7nc6PCO6ppTennP+364z3pdzvki80aD5KcBE54MP5Jy/SoHGun0QWOB26GosuP0MxsNH+xrg85+Nyd0twN6c85edJy9xmh1PyIWCgXdiNucMX7fXYHJ5DWbf+2LyexP2ltRuIpA6wWk5MOf8X8Rjzqt/gsnWNMwvfMLX+pWYbP8IT1qIPg0alDblUsx+XOz25Ejgcde7t2G2/2zi7Lb3Ybr+58CFOedzi/s2fda7nLZvwfhrtt/3S5gduDOl9NaSr4rx7cZkeCqmiz/o138Ze/vXVCy4eDomyzMxGbnFaXAGFqCajPldF2G6ea6v00if63NExfXZRJu0NpyqHnkM6Jtz/pD7ufN93Sf5396O6fFbsLerDsf2ONOJYwWexnjpOB/v6zHZneDj7PTfKemvipwfYbp/ALDCaXg1pj9Vqb/DafwdjNceJCqatgNvd9/p05gfoWTQbB/7BkwWPuk89hGMd8fknP/e9zzjfH/183LdG2v3BszP3uf3nw/c5rJZ3nc/5ltciPmHWuM/db05CrPvX8KCeAuwQMEQzPf8tdOxDbNNOzGbsBR7O/k3MZ0+0a9Z5UO8FwuCXYnJrPbIGzA+OsLHfKz/fbX/f3PO+Qb3JS4gKqcXY3Kts4GXEW80PRbTsV8APuJzuxzT3Z91Gh2L8civnG4dTo+HsMKEYRivL3ef6Cafh/bho32+CgYu8fWdhfHlakxHHI3x/UM+r3biyI63EHuodwI/dB77EaajT8Bk7sOYD/cc0JZzflNK6WM+h4ewPWF/otJzCGZvv06cXTbFafxT/6luk2W+Np8lDhLX+Wf/BqwsdIv8wguJA+P/2Nd8S875yZTS54BhOecPpJSuwHS7kqFn+b3HAR9v2G7p9N7EG4WnOH3myXd9PrxUAkk3YwuriKUihCrBV9R7OJExVVmlSsl06KKyDk8T/dsD/FF9iu8pW6+2naHEAXSv9etn4a+DzDlf6OM8jyiD708cBKgoZmdxv2v9urNzzhc25noccaBtOVedSbAXE+g+2Obx8vJ+mDM3mqjo2YMpsXOJDLDaDToLGuxwuug5vbGgkpzgCZgSfxx7leAVmFAf05inKn26aH3DkdZDlRHPNuY+x+dweUHvNxFnBij7rXVbWcxfveZqc9TcNhGKd04P4y3n34cod1UWXbxF49pZjbENxqLzV2AK4tRivPo+xX1UtaTzce7x78z0cX6CyNSo9aVsyVDpu1oIy9YvZXDV1qGg5w6iNeNep7We2ZQxrZ/4uE/xrEQccKkWSa2D2tvKNdrrNJfMiP/LtShlTbx8hY8NjN8kt1o3Xa9zL4b5/2f4d84uZFPza8q07tdcn/IayaMO/hMd73XalPcuZUnPavLfiZj8HuW0uYqQ4XcTbQv6eSKRoVOGo9RHJY31XGV1yu/d24OuKenZ5G1l8Cf4M8vxHWq9NfeSps21ko5sGVNjXGoradGVDXqXOr8sB+6glad7GpvGdAzRFis7obV9FeFgq42i5IFBRHuX5L/UXQfsQzG35+MB8euy4rmDiKyu1uYoH7faXvoS2clrCd3xOHBqznl8wz7JcVHVyTKiBbKpF5u81fxb0yaUMtpESfOZhJ1q6simTWrKe8lzpXw1ZfwYQs+VYzvwvUPoh3LefWg9g6Ucb3NMpTyWdrknG9mTbTrgD/QgB886vXpjm7QJuOw4vyhI1tN6l3qjHI8CZYfS300atYxRi9Xg66Y+78n2lpXJ/bFghvTcYkKP98RLJd1ehW1cTsVkZXExr6b+urdx/WjCt4GD+ezZQ+ilnvSlbKN0x0x69gVFb3r43NQVyxpzO6AfeqC59HHLujSu68mvaiPOvstEG0rTpzm7wQN65gTiIHbZiFK/lz5waZPKNZ6BJcp2P88zJQelHyo/TfrxoDVr0KC8R0lz6VHJVNO2luPNOefxh0Fb2a2edMKh1qjkF9mVxwn/rKm7Vfl5gHcI/VD693BoXS45V4JMyRKIg4Pl084k9PVdhHxpDbRH0aZS9xVKn7f0aeVXJaJNp8mL5X6j6ReX+5PSZ2rqt5Jnmvay6dP0aEPpAS6nLTqZVn13QDb99+cRvmoudLX2faWPPwMLCJQ6E2L9y/Vs7l/v4mC+GEPswdoIv0J+onhLtJHvW/r9qiwu10oVzOXetNTl0ok92djSto7DbMGxxB6jlO/Sf1T1j3Su5PZQ/nype3T/RT6/pg93BaYTxvj31aLc5Hvtf3Tfo2jdM2t8zb1N0/4e5KfTs08gm9Dct61u3KfcA5d7uUP524e03T3p0yZ6v9AFfyD4SvF5MhZdBItQapL3+99mYwR5GlOmQzFmPQeL9OLfud8/dxAZP5Xxn45Fde8nWjZOBxZ7tuEygJzzjSmlaY1xPoRlzzr8OSo3n+zXzNb9cs43AvRwD7BMg8r8J/p9jvc5n04ITUfOeW5KaX3jftN9jNemlP4Vy4bc6p9VcfI6v+/sggaLsQisnjMxW6XMZX7ddGCmP/NTOefPp5QeIUrzNc/jsTWYjUV1H6B1PbZiQlziKznn+3wO6wt6X+3jUZZT63I6lvGZ69cPcdpPLmg2u7j/cL9fc7zl/Idgiq0n3qK81te/HNsop/HwnPPHPcv0usb3Ke7zbv/udGCVf/diYKOPswvjR827kwiqiIdP8d+9Gxjt2ZNPYRmQI3wuP8UytkuxbNsU4NU55487rS8m3mAFIWNaP/HxkOJZO/z5o7CNe7kOi/0+B9bI5zC5kJnT/XO5FgdkreDl4cU4P0XIrdYNf8YQ5/VpWNbwc/4dyVapQ1pkurhfidmNaySPx/vcRcdFtBqtUp+Uc4JW/jvfeeQyYKuv/fqCJh1+X/08H1tHZTjaadVHB2hcPPd0H2/5vVLmmnq1J95+2j9fDLSX43ue9dbcS5q2rBWhI5tjOjAun9sBXoED61nSu9T5OE0v9r+VPN3T2DSmdzuNTiGq0TSu6T5WMBl6gFYe+DBmV0r5L3VXqdvBddzz8ICuv6Z47uv8WeXaTCvGDSaHiwi56iL0tOhR2qfJmA5+NbHGn6LVXokfmrzV/Nv04llNGW3iAM1fQEc2bVJT3kueE5pyeOA5cJD+KL8n2pTfLefdoTnSaoMOGlMpj6VdPoSNPMg2+ThbdFaDj6ZjVbz3+bpKdq7J/preQ6z3Ab1Rjsc/t4yXnmkLrfb5+fi6qc8PzK/4+xBMn8q+3O60lUxMo9VP60mXAEx3vf964ow+zaupvxY1rn9d4dvAwXzWxCH1JWEbwei8kR58wYLe9PC5SdNryrk19AO00lz6uHmPA9dBj34VmA55nX++nfBLSp02rbxXYaOmA9tKG0Grfi994AM2qVzjnPPnUkq7CPk66JlOo9LWTcf4ZzGhH58PzXscpEcLmWra1nK8Uw6TttM4tE7ocY1o5ZcPY1WC0t8lP0l3tzfmAa32Sv49HFqXn+9//54/cylxGHc7YZdafFqXm6uJiiM9S34jhA8ulD5v6dPKrxqN2deJfv0BXmzsN1r84sb+pPSZWvQbrTzTtJctPs3z2NCecA0H6+RS35WyORyrCr+2wU/lvm+mf5ZsPEyrzoRY/3I9W/avTpcmX8jHWel0ll8hP1G8tYhW33cm4fd3+u/KtRrEwXvT8xty05M/dsB/8Hm8x/cwVxN7jFK+S//xfFp98Xae358vdc/VWLX/Z5LtiZs+3HDCJ5QsTOyB77X/0X0vo3XPDD3sbXqwvz356T35BE07vxFa9Fq5jmXcQHu5Hv3tF7DdL4iXREVSRUVFRUVFRUVFRUVFRUVFRcXvH22/7wFUVFRUVFRUVFRUVFRUVFRUVLw0UANJFRUVFRUVFRUVFRUVFRUVFRWHhRpIqqioqKioqKj4HSCldEFK6eMv8jvfTCm943c1poqKioqKioqK3xYvlcO2KyoqKioqKir+3ZFSStiZkftf8OIXiZzzHcAd/973raioqKioqKj4faJWJFVUVFRUVFS8rJBSOjql9HRK6VvYK3UnpZTOTSk9lFJ6PKX0v1JKg/3apSmlz6eUfp1SejSldGxxj3tSSnNTSnenlI7s4TnvSyld75+/mVK6NqX0YEppsaqOkuF6H88viNcOk1I6PaV0X0ppdkrpzpTS+JRS75TSYyml6X7N/0gpXfW7pllFRUVFRUVFhVADSRUVFRUVFRUvR0wB/mfOeSr2KuFPAm/KOZ8G/Ar4b8W1nTnnk4HrgS/6764Dbso5TwO+C1x7GM8cD5wDvA3QK9gvwl7TeyLwHuBsgJRSH3/GO3LOpwPfAK7KOe8D3gfckFJ6E3AecOWLnn1FRUVFRUVFxW+I2tpWUVFRUVFR8XLEspzzw/75TCyQM8s63egLPFRce3Px8wv++Szg7f7528DnD+OZt3kL3VMppbH+u9cBN+ecu4HVKaV7/PfHAycBd/mYegFrAHLO81NK3wZ+DJyVc95zeFOuqKioqKioqPjtUQNJFRUVFRUVFS9H7Cg+J+CunPO7DnFtPsTnF4vdjWc+HxIwP+d81iH+fjKwhaIVrqKioqKioqLi/wVqa1tFRUVFRUXFyx0PA68tzj8alFI6rvj7pcVPVSo9CPyZf/5z4Je/4bPvBy5NKfVKKY0H3uC/fxoYnVI6y8fUJ6U01T+/HRiBVTNdl1Ia9hs+u6KioqKioqLiRaNWJFVUVFRUVFS8rJFz3pBSeh9wc0qpn//6k8Az/nl4SmkuVlGkqqXLgRtTSh8FNgCX/YaP/yHwH4CngOV4oCrnvMcP5L42pTQU89m+mFJah52v9Mac8wo/zPtLwHt/w+dXVFRUVFRUVLwopJx/mwrtioqKioqKior/f5FSWgqckXPe+PseS0VFRUVFRUXFHwJqa1tFRUVFRUVFRUVFRUVFRUVFxWGhViRVVFRUVFRUVFRUVFRUVFRUVBwWakVSRUVFRUVFRUVFRUVFRUVFRcVhoQaSKioqKioqKioqKioqKioqKioOCzWQVFFRUVFRUVFRUVFRUVFRUVFxWKiBpIqKioqKioqKioqKioqKioqKw0INJFVUVFRUVFRUVFRUVFRUVFRUHBZqIKmioqKioqKioqKioqKioqKi4rDwfwG5MAuLti0zDgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2fa6ea90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "scipy.cluster.hierarchy.set_link_color_palette(['m', 'c', 'y', 'k'])\n",
    "plt.figure(figsize=(20, 11))\n",
    "plt.title('Hierarchical Clustering Dendrogram')\n",
    "plt.xlabel('repo index')\n",
    "plt.ylabel('distance')\n",
    "\n",
    "dendrogram(\n",
    "    Z_norm,\n",
    "    #truncate_mode='lastp',\n",
    "   # leaf_font_size=10.,  # font size for the x axis labels\n",
    "   # labels=,\n",
    "    show_contracted=True,\n",
    "    orientation='top'\n",
    "   # above_threshold_color='y'\n",
    ")\n",
    "plt.axhline(y=.5, xmin=0, xmax=4, linewidth=4, color = '#d95f02')\n",
    "#plt.axhline()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2e154fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cut_off = [i for i in np.arange(0.1, 0.7, 0.02)]\n",
    "silhouette = {}\n",
    "dm = pearson_dist(name_fit_norm)\n",
    "for co in cut_off:\n",
    "    nodes = fcluster(Z_norm, co, criterion='distance')\n",
    "    silhouette[co] = silhouette_score(dm, nodes, metric='correlation')\n",
    "plt.figure()\n",
    "plt.plot(list(silhouette.keys()), list(silhouette.values()))\n",
    "plt.xlabel(\"Cut-off Distance\")\n",
    "plt.ylabel(\"Silhouette Score\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 356,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "1346"
      ]
     },
     "execution_count": 356,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hac_data_norm = tn_data\n",
    "hac_data_norm['pred_label'] = fcluster(Z_norm, 0.3, criterion='distance')\n",
    "test_hac_norm = hac_data_norm[hac_data_norm['repo'].isin(eval_table['repo'])]\n",
    "len(set(hac_data_norm['pred_label']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 357,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.948\n",
      "Completeness: 0.566\n",
      "Recall is 0.69\n",
      "Precision is 0.81\n"
     ]
    }
   ],
   "source": [
    "eval_ac_norm = eval_table[['repo', 'label']].merge(test_hac_norm[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_ac_norm['label'], eval_ac_norm['pred_label'])\n",
    "completeness = completeness_score(eval_ac_norm['label'], eval_ac_norm['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_ac_norm.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_ac_norm.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 358,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.989\n",
      "Completeness: 0.566\n",
      "Recall is 0.68\n",
      "Precision is 0.88\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "test_hac_norm = testset\n",
    "Z_test_norm = linkage(normalize(test_fit, axis=1, norm='l2'), metric = 'correlation')\n",
    "test_hac_norm['pred_label'] = fcluster(Z_test_norm, 0.3, criterion='distance')\n",
    "eval_ac_norm = eval_table[['repo', 'label']].merge(test_hac_norm[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_ac_norm['label'], eval_ac_norm['pred_label'])\n",
    "completeness = completeness_score(eval_ac_norm['label'], eval_ac_norm['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_ac_norm.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_ac_norm.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## KMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished calculating k = 50\n",
      "Finished calculating k = 100\n",
      "Finished calculating k = 150\n",
      "Finished calculating k = 200\n",
      "Finished calculating k = 250\n",
      "Finished calculating k = 300\n",
      "Finished calculating k = 350\n",
      "Finished calculating k = 400\n",
      "Finished calculating k = 450\n",
      "Finished calculating k = 500\n",
      "Finished calculating k = 550\n",
      "Finished calculating k = 600\n",
      "Finished calculating k = 650\n",
      "Finished calculating k = 700\n",
      "Finished calculating k = 750\n",
      "Finished calculating k = 800\n",
      "Finished calculating k = 850\n",
      "Finished calculating k = 900\n",
      "Finished calculating k = 950\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2f20be10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sse = {}\n",
    "for k in range(50, 1000, 50):\n",
    "    kmeans_norm = KMeans(n_clusters=k, max_iter=200).fit(name_fit_norm)\n",
    "    #data[\"clusters\"] = kmeans.labels_\n",
    "    #print(data[\"clusters\"])\n",
    "    sse[k] = kmeans_norm.inertia_ # Inertia: Sum of distances of samples to their closest cluster center\n",
    "    print(\"Finished calculating k = \" + str(k))\n",
    "plt.figure()\n",
    "plt.plot(list(sse.keys()), list(sse.values()))\n",
    "plt.xlabel(\"Number of cluster\")\n",
    "plt.ylabel(\"SSE\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 515,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "kmeans_data_norm = tn_data\n",
    "kmeans_norm = KMeans(n_clusters=150, random_state=122).fit(name_fit_norm)\n",
    "kmeans_data_norm['pred_label'] = kmeans_norm.labels_\n",
    "kmeans_test_norm = kmeans_data_norm[kmeans_data_norm['repo'].isin(eval_table['repo'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 520,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "836.3866817666603"
      ]
     },
     "execution_count": 520,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans_norm.inertia_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 516,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.789\n",
      "Completeness: 0.723\n",
      "Recall is 0.15\n",
      "Precision is 0.48\n"
     ]
    }
   ],
   "source": [
    "eval_kmeans_norm = eval_table[['repo', 'label']].merge(kmeans_test_norm[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_kmeans_norm['label'], eval_kmeans_norm['pred_label'])\n",
    "completeness = completeness_score(eval_kmeans_norm['label'], eval_kmeans_norm['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_kmeans_norm.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_kmeans_norm.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 369,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.856\n",
      "Completeness: 0.557\n",
      "Recall is 0.69\n",
      "Precision is 0.69\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "test_kmeans_norm = testset\n",
    "kmeans_test_norm = KMeans(n_clusters=77, max_iter=1000).fit(normalize(test_fit, axis=1, norm='l2'))\n",
    "test_kmeans_norm['pred_label'] = kmeans_test_norm.labels_\n",
    "\n",
    "eval_kmeans_norm = eval_table[['repo', 'label']].merge(test_kmeans_norm[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_kmeans_norm['label'], eval_kmeans_norm['pred_label'])\n",
    "completeness = completeness_score(eval_kmeans_norm['label'], eval_kmeans_norm['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_kmeans_norm.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_kmeans_norm.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 522,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1a33cc4e48>"
      ]
     },
     "execution_count": 522,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2c63be48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8, 8))\n",
    "plt.scatter(x, y, c=kmeans_data_norm['pred_label'])\n",
    "plt.ylabel('Component 2')\n",
    "plt.xlabel('Component 1')\n",
    "plt.title(\"Segregation of Topic Clusters\", )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Recommend similar Repos for a given one**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 446,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get the similarity matrix using Pearson Correlation\n",
    "Pearson_simMatrix = 1 - pearson_dist(name_fit_norm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 391,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getSimRepo(repo, simMatrix, k):\n",
    "    '''Return the top k most similar repos to the given one'''\n",
    "    simMatrix = pd.DataFrame(simMatrix, index=tn_data['repo'], columns=tn_data['repo'])\n",
    "    #repoid = tn_data[tn_data['repo'] == repo].index.tolist()[0]\n",
    "    rec = list(simMatrix.loc[repo, :].sort_values(ascending=False)[1:k].index)\n",
    "    return rec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 529,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['TheAlgorithms/Python',\n",
       " 'django/django',\n",
       " 'bitcoin/bitcoin',\n",
       " 'tornadoweb/tornado',\n",
       " 'Jahaja/psdash',\n",
       " 'shawnwun/RNNLG']"
      ]
     },
     "execution_count": 529,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getSimRepo('keon/algorithms', Pearson_simMatrix, 7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 573,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Tokutek/mysql-5.5',\n",
       " 'Tokutek/mariadb-5.5',\n",
       " 'facebook/mysql-5.6',\n",
       " 'therecluse26/PHP-Login',\n",
       " 'akopytov/sysbench',\n",
       " 'sbabic/swupdate']"
      ]
     },
     "execution_count": 573,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getSimRepo('twitter/mysql', Pearson_simMatrix, 7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 574,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['pytorch/pytorch',\n",
       " 'aciditeam/torch-models',\n",
       " 'jcjohnson/neural-style',\n",
       " 'keras-team/keras',\n",
       " 'tensorflow/models',\n",
       " 'fastai/fastai']"
      ]
     },
     "execution_count": 574,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getSimRepo('torch/nn', Pearson_simMatrix, 7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 575,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['markjaquith/WordPress',\n",
       " 'jumilla/wordpress-plus',\n",
       " 'dxw/wordpress',\n",
       " 'mhoofman/wordpress-heroku',\n",
       " 'owen2345/camaleon-cms',\n",
       " 'twilio/OpenVBX']"
      ]
     },
     "execution_count": 575,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getSimRepo('WordPress/WordPress', Pearson_simMatrix, 7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Including comments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 532,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "com_data = sub_data[['repo']]\n",
    "com_data['combine'] = sub_data['stemmed_name'] + sub_data['stemmed_comment']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 379,
   "metadata": {},
   "outputs": [],
   "source": [
    "buckets_length = []\n",
    "for i in filter(is_prime, range(500, 1800, 29)):\n",
    "    buckets_length.append(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 587\n",
      "Finished calculating k = 140 for bucket length = 587\n",
      "Finished calculating k = 180 for bucket length = 587\n",
      "Finished calculating k = 220 for bucket length = 587\n",
      "Finished calculating k = 260 for bucket length = 587\n",
      "Finished calculating k = 300 for bucket length = 587\n",
      "Finished calculating k = 340 for bucket length = 587\n",
      "Finished calculating k = 380 for bucket length = 587\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 761\n",
      "Finished calculating k = 140 for bucket length = 761\n",
      "Finished calculating k = 180 for bucket length = 761\n",
      "Finished calculating k = 220 for bucket length = 761\n",
      "Finished calculating k = 260 for bucket length = 761\n",
      "Finished calculating k = 300 for bucket length = 761\n",
      "Finished calculating k = 340 for bucket length = 761\n",
      "Finished calculating k = 380 for bucket length = 761\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 877\n",
      "Finished calculating k = 140 for bucket length = 877\n",
      "Finished calculating k = 180 for bucket length = 877\n",
      "Finished calculating k = 220 for bucket length = 877\n",
      "Finished calculating k = 260 for bucket length = 877\n",
      "Finished calculating k = 300 for bucket length = 877\n",
      "Finished calculating k = 340 for bucket length = 877\n",
      "Finished calculating k = 380 for bucket length = 877\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1051\n",
      "Finished calculating k = 140 for bucket length = 1051\n",
      "Finished calculating k = 180 for bucket length = 1051\n",
      "Finished calculating k = 220 for bucket length = 1051\n",
      "Finished calculating k = 260 for bucket length = 1051\n",
      "Finished calculating k = 300 for bucket length = 1051\n",
      "Finished calculating k = 340 for bucket length = 1051\n",
      "Finished calculating k = 380 for bucket length = 1051\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1109\n",
      "Finished calculating k = 140 for bucket length = 1109\n",
      "Finished calculating k = 180 for bucket length = 1109\n",
      "Finished calculating k = 220 for bucket length = 1109\n",
      "Finished calculating k = 260 for bucket length = 1109\n",
      "Finished calculating k = 300 for bucket length = 1109\n",
      "Finished calculating k = 340 for bucket length = 1109\n",
      "Finished calculating k = 380 for bucket length = 1109\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1283\n",
      "Finished calculating k = 140 for bucket length = 1283\n",
      "Finished calculating k = 180 for bucket length = 1283\n",
      "Finished calculating k = 220 for bucket length = 1283\n",
      "Finished calculating k = 260 for bucket length = 1283\n",
      "Finished calculating k = 300 for bucket length = 1283\n",
      "Finished calculating k = 340 for bucket length = 1283\n",
      "Finished calculating k = 380 for bucket length = 1283\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1399\n",
      "Finished calculating k = 140 for bucket length = 1399\n",
      "Finished calculating k = 180 for bucket length = 1399\n",
      "Finished calculating k = 220 for bucket length = 1399\n",
      "Finished calculating k = 260 for bucket length = 1399\n",
      "Finished calculating k = 300 for bucket length = 1399\n",
      "Finished calculating k = 340 for bucket length = 1399\n",
      "Finished calculating k = 380 for bucket length = 1399\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hashing finished...\n",
      "vectorization finished...\n",
      "Finished calculating k = 100 for bucket length = 1747\n",
      "Finished calculating k = 140 for bucket length = 1747\n",
      "Finished calculating k = 180 for bucket length = 1747\n",
      "Finished calculating k = 220 for bucket length = 1747\n",
      "Finished calculating k = 260 for bucket length = 1747\n",
      "Finished calculating k = 300 for bucket length = 1747\n",
      "Finished calculating k = 340 for bucket length = 1747\n",
      "Finished calculating k = 380 for bucket length = 1747\n"
     ]
    }
   ],
   "source": [
    "sse = {}\n",
    "for b_len in buckets_length:\n",
    "    com_data['hash2vec'] = com_data['combine'].apply(lambda x: hash2vec(x, b_len))\n",
    "    print(\"Hashing finished...\")\n",
    "    com_fit = vectorize(com_data['hash2vec'])\n",
    "    com_fit_norm = normalize(com_fit, axis=1, norm='l2')\n",
    "    print('vectorization finished...')\n",
    "    sse_b = {}\n",
    "    key = 'bucket' + str(b_len)\n",
    "    for k in range(100, 400, 40):\n",
    "        kmeans = KMeans(n_clusters=k, random_state=122, max_iter=200).fit(com_fit_norm)\n",
    "        #data[\"clusters\"] = kmeans.labels_\n",
    "        #print(data[\"clusters\"])\n",
    "        sse_b[k] = kmeans.inertia_ # Inertia: Sum of distances of samples to their closest cluster center\n",
    "        print(\"Finished calculating k = \" + str(k) + \" for bucket length = \" + str(b_len))\n",
    "    sse[key] = sse_b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a30a80d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "n_cluster = [i for i in range(100, 400, 40)]\n",
    "for n_bucket, record in sse.items():\n",
    "    plt.plot(n_cluster, list(record.values()), label=str(n_bucket))\n",
    "plt.xlabel(\"Number of Cluster\")\n",
    "plt.ylabel(\"SSE\")\n",
    "plt.legend(title='Length of Hash Table', loc='best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 533,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "com_data['hash2vec'] = com_data['combine'].apply(lambda x: hash2vec(x, 587))\n",
    "com_fit = vectorize(com_data['hash2vec'])\n",
    "com_fit_norm = normalize(com_fit, axis=1, norm='l2')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## HAC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 381,
   "metadata": {},
   "outputs": [],
   "source": [
    "Z_com_norm = linkage(com_fit_norm, metric = 'correlation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 387,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2ff419b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "scipy.cluster.hierarchy.set_link_color_palette(['m', 'c', 'y', 'k'])\n",
    "plt.figure(figsize=(20, 11))\n",
    "plt.title('Hierarchical Clustering Dendrogram')\n",
    "plt.xlabel('repo index')\n",
    "plt.ylabel('distance')\n",
    "\n",
    "dendrogram(\n",
    "    Z_com_norm,\n",
    "    #truncate_mode='lastp',\n",
    "   # leaf_font_size=10.,  # font size for the x axis labels\n",
    "   # labels=,\n",
    "    show_contracted=True,\n",
    "    orientation='top'\n",
    "   # above_threshold_color='y'\n",
    ")\n",
    "plt.axhline(y=.3, xmin=0, xmax=4, linewidth=4, color = '#d95f02')\n",
    "#plt.axhline()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a2fe78828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cut_off = [i for i in np.arange(0.1, 0.7, 0.05)]\n",
    "silhouette = {}\n",
    "dm = pearson_dist(com_fit_norm)\n",
    "for co in cut_off:\n",
    "    nodes = fcluster(Z_com_norm, co, criterion='distance')\n",
    "    silhouette[co] = silhouette_score(dm, nodes, metric='correlation')\n",
    "plt.figure()\n",
    "plt.plot(list(silhouette.keys()), list(silhouette.values()))\n",
    "plt.xlabel(\"Cut-off Distance\")\n",
    "plt.ylabel(\"Silhouette Score\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 383,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "367"
      ]
     },
     "execution_count": 383,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hac_com_data_norm = com_data\n",
    "hac_com_data_norm['pred_label'] = fcluster(Z_com_norm, 0.5, criterion='distance')\n",
    "test_hac_norm = hac_com_data_norm[hac_com_data_norm['repo'].isin(eval_table['repo'])]\n",
    "len(set(hac_com_data_norm['pred_label']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 384,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.161\n",
      "Completeness: 0.511\n",
      "Recall is 0.95\n",
      "Precision is 0.50\n"
     ]
    }
   ],
   "source": [
    "eval_com_ac_norm = eval_table[['repo', 'label']].merge(test_hac_norm[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_com_ac_norm['label'], eval_com_ac_norm['pred_label'])\n",
    "completeness = completeness_score(eval_com_ac_norm['label'], eval_com_ac_norm['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_com_ac_norm.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_com_ac_norm.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 536,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/wangjingci/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    }
   ],
   "source": [
    "kmeans_com_data_norm = com_data\n",
    "kmeans_norm = KMeans(n_clusters=150, random_state=122).fit(com_fit)\n",
    "kmeans_com_data_norm['pred_label'] = kmeans_norm.labels_\n",
    "kmeans_test_norm = kmeans_com_data_norm[kmeans_com_data_norm['repo'].isin(eval_table['repo'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 537,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Homogeneity: 0.510\n",
      "Completeness: 0.754\n",
      "Recall is 0.44\n",
      "Precision is 0.45\n"
     ]
    }
   ],
   "source": [
    "eval_com_kmeans_norm = eval_table[['repo', 'label']].merge(kmeans_test_norm[['repo', 'pred_label']], on='repo').drop_duplicates()\n",
    "h_score = homogeneity_score(eval_com_kmeans_norm['label'], eval_com_kmeans_norm['pred_label'])\n",
    "completeness = completeness_score(eval_com_kmeans_norm['label'], eval_com_kmeans_norm['pred_label'])\n",
    "print(\"Homogeneity: %0.3f\" % h_score)\n",
    "print(\"Completeness: %0.3f\" % completeness)\n",
    "Recall = []\n",
    "Precision = []\n",
    "for name, group in eval_com_kmeans_norm.groupby('label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['pred_label']).most_common()[0][1]\n",
    "        recall = purity / group.shape[0]\n",
    "        Recall.append(recall)\n",
    "print('Recall is %.2f' %(sum(Recall)/len(Recall)))\n",
    "for name, group in eval_com_kmeans_norm.groupby('pred_label'):\n",
    "    if group.shape[0] > 1:\n",
    "        purity = Counter(group['label']).most_common()[0][1]\n",
    "        precision = purity / group.shape[0]\n",
    "        Precision.append(precision)\n",
    "print('Precision is %.2f' %(sum(Precision)/len(Precision)))\n",
    "      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 390,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['riot/riot',\n",
       " 'infernojs/inferno',\n",
       " 'vuejs/vue',\n",
       " 'jsdoc3/jsdoc',\n",
       " 'yogthos/Selmer',\n",
       " 'Meituan-Dianping/mpvue']"
      ]
     },
     "execution_count": 390,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cosine_simMatrix = 1 - cosine_similarity(com_fit_norm)\n",
    "getSimRepo('nfl/react-helmet', Pearson_simMatrix, 7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}