editor_pca_factors.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import csv, re\n",
    "from collections import Counter, OrderedDict\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.backends.backend_pdf import PdfPages\n",
    "from IPython.core.debugger import Tracer; debug_here = Tracer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/ipykernel/__main__.py:23: 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"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.2\n",
      "1 206\n",
      "2 127\n",
      "3 133\n",
      "4 46\n",
      "5 164\n",
      "6 95\n",
      "7 218\n",
      "8 127\n",
      "9 105\n",
      "10 198\n",
      "neutral 499\n",
      "\n",
      "0.6\n",
      "1 130\n",
      "2 96\n",
      "3 97\n",
      "4 20\n",
      "5 105\n",
      "6 75\n",
      "7 143\n",
      "8 108\n",
      "9 72\n",
      "10 136\n",
      "neutral 936\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Talk page usernames\n",
    "talk_csvpath = '/home/michael/school/research/wp/wikipedia/data/talk/ipc_utf8_talkpages.csv'\n",
    "talk_counter = Counter()\n",
    "edpercent = 5\n",
    "\n",
    "with open(talk_csvpath, 'r') as csvfile:\n",
    "    reader = csv.reader(csvfile, delimiter=',')\n",
    "    for row in reader:\n",
    "        talk_counter[row[3]] += 1\n",
    "\n",
    "talk_eds = set(talk_counter.keys())\n",
    "\n",
    "# Create factor matrix of just editors on the talk pages\n",
    "cols = [\"Factor{:d}\".format(i) for i in range(1, 11)]\n",
    "factors = pd.read_csv('/home/michael/school/research/wp/wikipedia/data/editor_citation_matrix_factors.csv', index_col=0)[cols]\n",
    "factors_talk = factors[factors.index.isin(talk_eds)]\n",
    "# print(len(factors_talk))\n",
    "\n",
    "# Assign talk editors to their highest factors above certain threshold\n",
    "# thresholds = np.linspace(0.0, 2.0, 21)\n",
    "thresholds = [0.2, 0.6]\n",
    "assigned_factors = {}\n",
    "factors_talk['assigned_factor'] = factors_talk.idxmax(axis=1)\n",
    "editors = {}\n",
    "for t in thresholds:\n",
    "    editors[t] = {}\n",
    "    \n",
    "    assigned_factors[t] = factors_talk.copy()\n",
    "    \n",
    "    for name, row in assigned_factors[t].iterrows():\n",
    "        if max(row[:10]) < t:\n",
    "            assigned_factors[t].set_value(name, 'assigned_factor', 0)\n",
    "            \n",
    "    print(t)\n",
    "    for i in range(1,11):\n",
    "        editors[t][i] = assigned_factors[t][assigned_factors[t]['assigned_factor']=='Factor{:d}'.format(i)].index.tolist()\n",
    "        print(i, len(editors[t][i]))\n",
    "    \n",
    "    # neutral\n",
    "    editors[t][0] = assigned_factors[t][assigned_factors[t]['assigned_factor']==0].index.tolist()\n",
    "    print('neutral', len(editors[t][0]))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# CITATION WINDOW DATA--each editor assigned to one factor, editors selected by threshold\n",
    "# Gather article text from editors\n",
    "citewindow_csvpath = '/home/michael/school/research/wp/wp_articles/editor_citewindow_50.csv'\n",
    "art_data = pd.read_csv(citewindow_csvpath)\n",
    "art_txt = {} # art_txt[editor] = all_text\n",
    "for ed in list(talk_eds):\n",
    "    art_txt[ed] = art_data[art_data['editor']==ed].values[0][1]\n",
    "\n",
    "# Build up factor training text from top quantiles\n",
    "for t in thresholds:\n",
    "    with open('/home/michael/school/research/wp/wp_articles/editor_citewindow_factor{:1.1f}.csv'.format(t), 'w') as f:\n",
    "        w = csv.writer(f)\n",
    "        w.writerow(['factor', 'editor', 'text'])\n",
    "        for i in range(11):\n",
    "            for ed in editors[t][i]:\n",
    "                if not (isinstance(art_txt[ed], float) and np.isnan(art_txt[ed])):\n",
    "                    w.writerow([i, ed, art_txt[ed]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# ARTICLE CONTRIBUTION DATA--each editor assigned to one factor, editors selected by threshold\n",
    "# Gather article text from editors\n",
    "art_csvpath = '/home/michael/school/research/wp/wikipedia/data/ipc_article_diffs.csv'\n",
    "art_data = pd.read_csv(art_csvpath)\n",
    "art_txt = {} # art_txt[editor] = all_text\n",
    "for ed in list(talk_eds):\n",
    "    #for _, row in art_data[art_data['username']==ed].iterrows():\n",
    "    #    debug_here()\n",
    "    #    print(row['post_text'])\n",
    "    art_txt[ed] = \" \".join([str(row['additions']) for _, row in art_data[art_data['editor']==ed].iterrows()])\n",
    "\n",
    "# print(len(art_txt))\n",
    "\n",
    "# Build dict with top and bottom quantile of each factor\n",
    "# factor_eds = {i: {} for i in range(1,11)}\n",
    "# div = 100/edpercent\n",
    "# for i in range(1, 11):\n",
    "#     factor_eds[i]['top'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=False)[:round(len(factors_talk)/div)]\n",
    "#     factor_eds[i]['bottom'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=True)[:round(len(factors_talk)/div)]\n",
    "\n",
    "# Build up factor training text from top quantiles\n",
    "for t in thresholds:\n",
    "    with open('/home/michael/school/research/wp/wikipedia/data/editortext_factors_{:1.1f}.csv'.format(t), 'w') as f:\n",
    "        w = csv.writer(f)\n",
    "        w.writerow(['factor', 'editor', 'additions'])\n",
    "        for i in range(11):\n",
    "            #for ed in factor_eds[i]['top'].index:\n",
    "            #    if str(ed) not in art_txt.keys():\n",
    "            #        print(ed)\n",
    "        #     factor_txt[i] = ' '.join([art_txt[ed] for ed in factor_eds[i]['top'].index])\n",
    "            for ed in editors[t][i]:\n",
    "                w.writerow([i, ed, art_txt[ed]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Add editor factor assignments to article additions\n",
    "article_adds = pd.read_csv('/home/michael/school/research/wp/wikipedia/data/ipc_article_diffs.csv')\n",
    "\n",
    "for t in thresholds:\n",
    "    article_adds_factors = article_adds.copy()\n",
    "    for i, row in enumerate(article_adds.values):\n",
    "        ed = row[-2]\n",
    "        if ed in assigned_factors[t].index:\n",
    "            fac = assigned_factors[t].loc[row[-2], 'assigned_factor']\n",
    "            if fac != 0:\n",
    "                fac = int(fac.split('Factor')[1])\n",
    "            article_adds_factors.set_value(i, 'factor', fac)\n",
    "    \n",
    "    criterion = article_adds_factors['factor'].map(lambda x: not np.isnan(x))\n",
    "    article_adds_factors = article_adds_factors[criterion]\n",
    "    article_adds_factors.to_csv('/home/michael/school/research/wp/wikipedia/data/article_diffs_factors_{:1.1f}.csv'.format(t), index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEACAYAAAC9Gb03AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSWrc/wNwoZYkVfTWmwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1d668319e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "# Check means, variance of PCA factor loadings\n",
    "means = factors_talk.mean(axis=1)\n",
    "plt.hist(means.values, bins=1000)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAEACAYAAABYq7oeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUovhIElqMRwkSS3/H7oEa0EOf3E4AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1d66851438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "var = factors_talk.var(axis=1)\n",
    "plt.hist(var.values, bins=1000)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242\n",
      "237\n",
      "165\n",
      "74\n",
      "203\n",
      "105\n",
      "292\n",
      "181\n",
      "119\n",
      "300\n"
     ]
    }
   ],
   "source": [
    "# Number of editors with highest value in certain factors\n",
    "for i in range(1,11):\n",
    "    print(len(pca_editors[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# ARTICLE CONTRIBUTION DATA--each editor assigned to one factor--by percentage\n",
    "# Gather article text from editors\n",
    "art_csvpath = '/home/michael/school/research/wp/wikipedia/data/ipc_article_diffs.csv'\n",
    "art_data = pd.read_csv(art_csvpath)\n",
    "art_txt = {} # art_txt[editor] = all_text\n",
    "for ed in list(talk_eds):\n",
    "    #for _, row in art_data[art_data['username']==ed].iterrows():\n",
    "    #    debug_here()\n",
    "    #    print(row['post_text'])\n",
    "    art_txt[ed] = \" \".join([str(row['additions']) for _, row in art_data[art_data['editor']==ed].iterrows()])\n",
    "\n",
    "# print(len(art_txt))\n",
    "\n",
    "# Build dict with top and bottom quantile of each factor\n",
    "factor_eds = {i: {} for i in range(1,11)}\n",
    "div = 100/edpercent\n",
    "for i in range(1, 11):\n",
    "    factor_eds[i]['top'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=False)[:round(len(factors_talk)/div)]\n",
    "    factor_eds[i]['bottom'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=True)[:round(len(factors_talk)/div)]\n",
    "\n",
    "# Build up factor training text from top quantiles\n",
    "# factor_txt = {}\n",
    "with open('/home/michael/school/research/wp/wikipedia/data/top{:d}percent_article_adds.csv'.format(edpercent), 'w') as f:\n",
    "    w = csv.writer(f)\n",
    "    w.writerow(['factor', 'editor', 'text'])\n",
    "    for i in range(1, 5):\n",
    "        #for ed in factor_eds[i]['top'].index:\n",
    "        #    if str(ed) not in art_txt.keys():\n",
    "        #        print(ed)\n",
    "    #     factor_txt[i] = ' '.join([art_txt[ed] for ed in factor_eds[i]['top'].index])\n",
    "        for ed in factor_eds[i]['top'].index:\n",
    "            w.writerow([i, ed, art_txt[ed]])\n",
    "\n",
    "# Write out factor texts\n",
    "# data = pd.DataFrame(factor_txt, index=['text']).transpose()\n",
    "# data['factor'] = data.index\n",
    "# data = data[data.columns.tolist()[-1:] + data.columns.tolist()[:-1]]\n",
    "# data = pd.DataFrame(outlines)\n",
    "# data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# ARTICLE CONTRIBUTION DATA\n",
    "# Gather article text from editors\n",
    "art_csvpath = '/home/michael/school/research/wp/wikipedia/data/ipc_article_diffs.csv'\n",
    "art_data = pd.read_csv(art_csvpath)\n",
    "art_txt = {} # art_txt[editor] = all_text\n",
    "for ed in list(talk_eds):\n",
    "    #for _, row in art_data[art_data['username']==ed].iterrows():\n",
    "    #    debug_here()\n",
    "    #    print(row['post_text'])\n",
    "    art_txt[ed] = \" \".join([str(row['additions']) for _, row in art_data[art_data['editor']==ed].iterrows()])\n",
    "\n",
    "# print(len(art_txt))\n",
    "\n",
    "# Build dict with top and bottom quantile of each factor\n",
    "factor_eds = {i: {} for i in range(1,11)}\n",
    "div = 100/edpercent\n",
    "for i in range(1, 11):\n",
    "    factor_eds[i]['top'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=False)[:round(len(factors_talk)/div)]\n",
    "    factor_eds[i]['bottom'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=True)[:round(len(factors_talk)/div)]\n",
    "\n",
    "# Build up factor training text from top quantiles\n",
    "# factor_txt = {}\n",
    "with open('/home/michael/school/research/wp/wikipedia/data/top{:d}percent_article_adds.csv'.format(edpercent), 'w') as f:\n",
    "    w = csv.writer(f)\n",
    "    w.writerow(['factor', 'editor', 'text'])\n",
    "    for i in range(1, 5):\n",
    "        #for ed in factor_eds[i]['top'].index:\n",
    "        #    if str(ed) not in art_txt.keys():\n",
    "        #        print(ed)\n",
    "    #     factor_txt[i] = ' '.join([art_txt[ed] for ed in factor_eds[i]['top'].index])\n",
    "        for ed in factor_eds[i]['top'].index:\n",
    "            w.writerow([i, ed, art_txt[ed]])\n",
    "\n",
    "# Write out factor texts\n",
    "# data = pd.DataFrame(factor_txt, index=['text']).transpose()\n",
    "# data['factor'] = data.index\n",
    "# data = data[data.columns.tolist()[-1:] + data.columns.tolist()[:-1]]\n",
    "# data = pd.DataFrame(outlines)\n",
    "# data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "data.to_csv('/home/michael/school/research/wp/wikipedia/data/top{:d}percent_article_adds.csv'.format(edpercent), index=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAEZCAYAAABvpam5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2ebb81a518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "# Form clusters of editors based on a threshold\n",
    "\n",
    "factor_clusters = {}\n",
    "lo = 0\n",
    "hi = 50\n",
    "nvals = 251\n",
    "\n",
    "for t in np.linspace(lo, hi, nvals):\n",
    "#     print(\"Threshold: {:f}\".format(t))\n",
    "    factor_clusters[t] = {}\n",
    "    \n",
    "    for i in range(1, 11):\n",
    "#         print(\"\\tFactor {:d}\".format(i), end=\": \")\n",
    "        factor_clusters[t][i] = factors_talk[factors_talk['Factor{:d}'.format(i)] >= t].index\n",
    "#         print(len(factor_clusters[t][i]))\n",
    "    \n",
    "#     print()\n",
    "\n",
    "# Get #editors clustered at all depending on threshold want data (threshold, #editors clustered)\n",
    "total_clustered = [[] for _ in range(2)]\n",
    "for t in np.linspace(lo, hi, nvals):\n",
    "    total_clustered[0].append(t)\n",
    "    clustered = set()\n",
    "    for i in range(1,11):\n",
    "        factor_cluster = {name for name in factor_clusters[t][i]}\n",
    "        clustered |= factor_cluster\n",
    "    total_clustered[1].append(len(clustered))\n",
    "\n",
    "# print(total_clustered[1][:10])\n",
    "# print(total_clustered[0][:10])\n",
    "\n",
    "# Talk editors clustering by threshold, factors each plotted\n",
    "\n",
    "plt.plot(total_clustered[0], total_clustered[1], '.')\n",
    "plt.title(\"Talk editors in at least one cluster by threshold\")\n",
    "plt.xlabel('Threshold')\n",
    "plt.xticks(range(0, 51, 2))\n",
    "plt.axis([0, 20, 0, 2000])\n",
    "plt.yticks(range(0, 2001, 200))\n",
    "plt.ylabel('Number of editors clustered')\n",
    "plt.savefig('/home/michael/school/research/wp/pca_total_editors_classified.png', dpi=400)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAEZCAYAAABvpam5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2ebd8fe7b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "# Form clusters of editors based on a threshold--plot all factors\n",
    "\n",
    "factor_clusters = {}\n",
    "lo = 0\n",
    "hi = 50\n",
    "nvals = 251\n",
    "\n",
    "for t in np.linspace(lo, hi, nvals):\n",
    "#     print(\"Threshold: {:f}\".format(t))\n",
    "    factor_clusters[t] = {}\n",
    "    \n",
    "    for i in range(1, 11):\n",
    "#         print(\"\\tFactor {:d}\".format(i), end=\": \")\n",
    "        factor_clusters[t][i] = factors_talk[factors_talk['Factor{:d}'.format(i)] >= t].index\n",
    "#         print(len(factor_clusters[t][i]))\n",
    "    \n",
    "#     print()\n",
    "\n",
    "# Get #editors clustered at all depending on threshold want data (threshold, #editors clustered)\n",
    "thresholds = np.linspace(lo,hi,nvals)\n",
    "nclustered = {} # nclustered[factor]\n",
    "for i in range(1,11):\n",
    "    nclustered[i] = []\n",
    "    for t in thresholds:\n",
    "        factor_cluster = {name for name in factor_clusters[t][i]}\n",
    "        nclustered[i].append(len(factor_cluster))\n",
    "\n",
    "#     print(i)\n",
    "#     print(nclustered[i])\n",
    "\n",
    "# plt.figure(figsize=(8,6), dpi=400)\n",
    "# plt.plot(thresholds, *[nclustered[i] for i in range(1,11)], '-')\n",
    "for i in range(1,11):\n",
    "    plt.plot(thresholds, nclustered[i])\n",
    "plt.title(\"Talk editors in factor clusters\")\n",
    "plt.xlabel('Threshold')\n",
    "plt.xticks(range(0, 51, 2))\n",
    "plt.axis([0, 20, 0, 1000])\n",
    "plt.yticks(range(0, 1001, 200))\n",
    "plt.ylabel('Number of editors clustered')\n",
    "plt.savefig('/home/michael/school/research/wp/pca_editors_by_threshold_factor.png', dpi=400)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY8AAAEZCAYAAABvpam5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XmcXFWd9/HPNwvIEkJACUIgYRMhURZFAkhoQfaRODo6\n4MIyPojigug4IDOS+DjKooAgzKAjIPDIroPgoEQgLeiEnQAhBIIIhABBlpAGFUjye/44p5KbSlV3\nVXdt6Xzfr1e/+t5zb93zq+rq+tU5595zFRGYmZnVY0i7AzAzs1WPk4eZmdXNycPMzOrm5GFmZnVz\n8jAzs7o5eZiZWd2cPFpI0pqSlkraJK9fLumkJtSzraQ3C+s3S/pYo+upIY5tJL3U6npz3Su8Bqsr\nSftLmtvuOCqR9Kyk3Zt07FMk/bgZxy6rp9/vs77+Ns36fGgUJ486SeqRtCj/LJH0l0LZYTUcolUX\n1iyrJyL2iYirASQdI+m3LQkgYm5EbNCMY0uaIekTfYXQjLoLMbTkA6oBBvw6NPODfqA6IEEO5PVd\nZS+0G9buAFY1ETGitCzpceAzETG9jkOo8VHVrd9vWElDI2JJI4OxwU2SAKJ5VyQLv6dbzi2PgRFl\nyUDS7pJul/SypKclnSmpz9dZ0khJt0k6rcr2UZIuzt8An5R0cmHbUElnS3pB0qPAvmWPnSHpE5J2\nBH4AdOXW0jOFY18m6XlJf5T09cJjj8ndXufmLqgTclP9NkkLJS2Q9NMqMZd3n82QdHL+/YqkX0ka\nWeWxb5V0Q47pBUnXShqdt30f2AX4SW7xfa+G17e3129bSdMlvVh6PpLWKWz/pqRncswPSdpD0mTg\nq8AROYbbq9T7Lkm35vfDTEkHFLZdLuksSb/Jx7hN0maF7RPya/9SrndyL89vw8Lze1HS5RX2WaHb\ntBDDSXl5tKRf51hfkHRTLr8K2AiYluP8Yi7fs/Bev7vYMsl/42/l1+U14O1VQt9D0sO5vh9JGp4f\nP1fSPmWxL5S0bdlz2gD4BbCllvcAjMqb187v60X5tX934XHPSvqapFnAK7lss/w++7OkxyQdU9h/\nd0n35vfAM5K+s2IYOlLSvPz++efChrdIOi8/5ilJp0saWumFkPS+HOcrki4F1qjymnWGiPBPP3+A\nPwF7l5W9F3hPXt4CeBT4bF5fE1gKbJLXLwdOAt4G3At8o5e6biB98K8JjAbuAT6dt30FuD+Xbwjc\nBrxReOwM4BN5+RhgWtmxrwKuBNYCtgIeBw4r7P8G8E+kRPkW0j/rVwvPabcqMW9bIY6HgXG5rj8A\nJ1d57EbAh0j/QCOA/wYuKzvWYb28XuV19/b6bQt0AUNzvf8LfDdvezfwR+CteX0cMDYvnwL8uJcY\n1gSezH+focB+QE/h8ZcDzwE75O1XAxfmbSOAZ4BD8/p7gBeALavUdTNwcX7cMOD9uXx/4NFCPEvI\n77/iezAvnwmclf/Oy46Rtz1b/DsDY3M8H8jrBwDPAyMLf5/HgK3zcxtSIeZn89+h9L69sxDLN4Gf\nFvb9OHBHlee+7DkWyk4BXgX2zs/nTGB6Wd135LrXJH2RfgD45xzv1sATwJ55/3uBj+bldYBdCu+d\npcAPSe/V9wKvA+Py9tOB3wGjSP/nd5L/z8v+Nm/Jf+9jcv2fAN4svR6d+OOWR4NFxN0RcU9e/hNw\nAbBXLw8ZC9wK/CQiTqm0g6TNgT2Br0XE6xGxgPRmPTTv8jHgjIhYEBEvkt6wNZG0BvAR4F8i4q8R\n8UfSh+ynC7s9HhEXRvI30pt6nKSNczwzaq0P+K+IeCIi/gpcA+xYaaeIeD4iro+INyKiBziNlV/H\nmroAJY2l8ut3WK7rkYjojoglEfE8cHahrsWkf+wJSt0bT0TEkzU+1z2BpRHxg3zsacBvgX8s7HNV\nRNwfqdvkMpa/Hn8PPBgRV+QY7wF+BXy0wvMbB+wBfD4ieiJicUT8vtrL0Uu8bwKbkD74Kh2j+Ngj\ngJ9H7rKNiN8As0kJsuQnEfFYfu5Lq9T5g8L79hTy3wS4BPiwpDXz+qeBS3uJvZJbIuKWSJ/Ol5KS\ndNGZue7XgfcDa0bE93O8jwE/Zfn/2BvAOyRtEBGvRcRdheME6UvQGxFxNzCH9KUDUhI4OSJejog/\nA//Oiv9bJZOAv0bEj3L9l5GSWcdy8mgwSdvl7pbnJL1C+gb11l4eMpn0zeWiXvYZS/qm/ufchfEy\n6QN+o7x9E2BeYf9aP9wANiZ9KJQ/ftPC+jxW9BXSt6/7cjP7k3XU91xh+S/AupV2krSupAtyF9NC\n4EZ6fx17szmVX7+35breLukqpW7GhcBPSnVFxGzgROA7wAJJl0p6W431bgI8VVZW/tpWez3GAnvl\neEsxf4TK3T+bAc9HxF9qjKuafyd9I58u6VFJx/ey71jg02XxvacsvvL3TSVPF5afJL1m5AR9LymB\nvJXUgrii9qcC9P1eK9Y9Ftii7PkcT2qZQEqWOwCP5i65YpJcEhEvV6lrY1Z8D5T//UveXhZPad+O\n5eTReP9FaopvEREjgW/T+7e9H5K6Sa4vfMsqNw/oiYgN8s+oiFg/InbJ258lfYCUjO2lvvKBxedI\nyWvzQtnmwPxqj4mIZyPiMxHxduA44EJJY3qpsz9OJP2TvSci1id9oy2+jvUMkPb1+n2P1MWxfa7r\n/xTriohLI2IPYEtSEvr3GmN4hhVfV1j5te0t5hvLYl4vIr5aZd+NJK3dxzHfILUuivttXFrIrZav\nRMQ4Ugvn3yTtVtpcoc7/KotvREScU9inlr9R+fv2mcL6JaRv6YcCN0fEC1WO0d/B8uLj5gEPlz2f\nkRHxD7CsdXoo6QvHD4FfSKrlhKPnWPH/cSyV//7PAuX/Q+XvnY7i5NF46wKvRMRfJY0Hju5j/4iI\no0n/NL/M3UjlOzwB3J4H29ZVsrWkPfIuVwHHS9o4f0v7evkxChYAm5Xe+BHxBmk84buS1pa0FSkh\nVO0ikPRxSaVvmK+Q/gmrna3S37PLRpC+wS3Kz+nfKjyPLfs4Ruksnyfo/fUbQUoer+YuwmUf0Lkl\nOSn/XV4H/kpKtqUYtuil/tuAIZK+rHRSw76kkxmu7CNugGuBnfJrPUzSGpJ2lbRN+Y75+d0KnCtp\nPUnDJe1ZYb8gdYV8UtIQSYcApeSApA9JKj2fHlKXXem5PseKr/fFwMck7Z2PtVZe3oj6fLnwvj2B\nFVsX15C6kz5HSiTVLCAlz3V62Qd6fy/+HkDScUqD88OUTnbYKZd/OndZBbCI9LqUkk9vx70cmCJp\ng/zanETl/61bgbdI+mx+rxzG8q6vjuTkMTCVvvEcDxwtaRHpG0p5UzuqLB8JvAxcU+UbzWHA+qT+\n1BfzcUv/qOeSPqgeIg1Uln84Fev5DWkg8HlJpeb050j/AE8CN5EGgVc6W6dgN+Ce/ByvBI6OiGer\n7Fvt+fbl+6RveS+S/rH+p2z7WaQznV6UdGoNdff2+p1MGp9YCPyc9KFVshZwBvBn0jfGdUhdkeRj\nrJO7OVYaY8h96X9HGpN6MT+njxfGTKq+HhGxkDSgehTpW+nTpFZstW+7h5EGbOfm/T9XZb8vkb7J\nvwQcAlxf2LYdqctqEdANfC8i7sjbvkv6gvGSpGPzeN5HgW+RBs7/BHyZ5Z8ptfytg/QaTgceIZ30\nsezMuYh4Nce3CXBd1YNE3J+3P5njW7+X+iotExGLgYOA3Un/BwuA/yD9vSH9HR/JXdHfAT4Wy0/v\nLX+uxfWTSWNBD5G64W4rPsdC/X8jjXN9gfS3OZBennMnUEqkTTp46sq4hNRvuJTUzD1H6VS6K0lN\nuCdI/1Cl0+XOIb1wrwFHRsTMXH4E8K+kP8x3IqK3byJmNggonRL7toj4bLtjsRU1O3lsDGwcETMl\nrUsaC5hM+jb1YkScLukEYFREnCjpQOCLEXGwpF2BsyNiYk42dwM7k74h3wPsXEo4Zjb45BMTZgKT\n81lM1kGa2m0VEc+VWg65CfowaVBoMqnPlPy7dPHTZHLfZm4uj1S6MGx/0rUJr+Tm/DTSeeVmNghJ\n+gKpK+xKJ47O1LLpSZTORd8RuB0YHelceyLiuZwgIJ1dUzy97+lcVl4+n8qnu5nZIBAR5wHntTsO\nq64lA+a5y+oa4LjcAultgGmFhzY1MDMz65emtzzymUPXAJdGxC9z8QJJoyNiQR4XeT6Xz2fF877H\n5LL5pOkjiuUrTUYoaZWdodLMrJ0ioq4v661oeVwIzI6Iswtl15FOTSX//mWh/HAASROBhbl760Zg\nX6XJA0eRzpW/sVJl0QFzvkyZMsUxdFAcnRBDp8TRCTF0ShydEEOnxNEfTW155IuwPgk8KOk+UvfU\nSaR5iq6S9E+kc6o/DhARN0g6SNJjpFN1j8rlL0v6NumMqwC+FWng3MzM2qCpySMi/kCaIbKSD1Z5\nzBerlP+UNFGZmZm1ma8wb4Kurq52h9ARMUBnxNEJMUBnxNEJMUBnxNEJMUDnxFGvpl4k2GqSYjA9\nHzOzVpBEdOCAuZmZDTKDLnn09LQ7AjOzwW/QJY8993QCMTNrtkGXPGbPhoceancUZmaD26BLHttv\nD+PHtzsKM7PBbdCdbbVoUTBiRLsjMTNbdfTnbKtBlzwG0/MxM2sFn6prZmYt4eRhZmZ1c/IwM7O6\nOXmYmVndnDzMzKxuTU0eki6QtEDSA4WyHSTNkHSfpDsl7VLYdo6kuZJmStqxUH6EpEclPSLp8GbG\nbGZmfWt2y+MiYP+ystOBKRGxEzAlryPpIGCriNgGOAY4P5ePAk4GdgF2BaZIGtnkuM3MrBdNTR4R\n8Xvg5bLipUDpw3990v3JAQ4BLsmPuwMYKWk0KflMi4hX8t0DpwEHNDNuMzPrXVPvJFjF8cCNks4A\nBOyeyzcF5hX2ezqXlZfPz2VmZtYm7UgenweOi4hrJf0DcCGwb4X96rrasWTq1KnLlru6ulbZu3SZ\nmTVLd3c33d3dAzpG06cnkTQWuD4i3p3XF0bE+oXtCyNifUnnA9Mj4spcPgfYC/gA0BURn8vlK+xX\nVpenJzEzq1OnTk8iVmxFzJe0F4CkfYC5ufw64PBcPhFYGBELgBuBfSWNzIPn++YyMzNrk6Z2W0m6\nDOgCNpT0FOnsqqOBcyQNBf4GfBYgIm6QdJCkx4DXgKNy+cuSvg3cDQTwrTxwbmZmbeJZdc3MVnOd\n2m1lZmaDjJOHmZnVzcnDzMzq5uRhZmZ1c/IwM7O6OXmYmVndnDzMzKxuTh5mZlY3Jw8zM6ubk4eZ\nmdXNycPMzOrm5GFmZnVz8jAzs7o1NXlIukDSAkkPlJV/SdLDkh6UdGqh/BuS5uZt+xXKD5A0R9Kj\nkk5oZsxmZta3pk7JLun9wKvAJYU7CXYBJwEHRcRiSW+NiBckbQdcBuwCjAFuArYh3UjqUWAf4Bng\nLuDQiJhToT5PyW5mVqf+TMne1JtBRcTv821oiz4PnBoRi/M+L+TyycAVufwJSXOB95GSx9yIeBJA\n0hV535WSh5mZtUbV5CHpI709MCJ+0c863wFMkvRd4K/AP0fEPcCmwIzCfvNzmYB5hfKnSUnFzMza\npLeWx4fy742A3YFb8voHgP8F+ps8hgGjImKipF2Aq4Et+3mslUydOnXZcldXF11dXY06tJnZoNDd\n3U13d/eAjtHnmIekacAREfFsXn878NOI2L+mClK31fWFMY8bgNMi4nd5fS4wkXRvcyLi1Fz+G9I9\nzwVMjYgDcvmJabc4rUJdHvMwM6tTs25Du1kpcWQLgM3riSv/lFwL7A0g6R3AGhHxInAd8I+S1pC0\nBbA1cCdpgHxrSWMlrQEcmvc1M7M2qWXA/GZJNwKX5/V/JJ0J1SdJlwFdwIaSniK1JC4ELpL0IPA6\ncDhARMyWdBUwG3gTODY3I5ZI+iIwjZTsLoiIh2t8fmZm1gQ1naor6e+BSXn11oj476ZG1U/utjIz\nq18zT9W9F+iJiJskrS1pRET01B+imZkNBn2OeUg6GrgG+FEu2pQ0bmFmZqupWgbMvwDsASwCiIi5\npNN3zcxsNVVL8ng9It4orUgaBnhgwcxsNVZL8vidpJOAtSTtS7qo7/rmhmVmZp2slosEhwCfAfYj\nXa9xI/CTTjytyWdbmZnVrz9nW/WaPCQNJc2I+8mBBtcKTh5mZvVr+BXmEbEEKF3ZbWZmBtR2ncfj\nwB8kXQe8ViqMiDObFpWZmXW0WpLHH/PPEGBEc8MxM7NVQc13EpS0dkT8pcnxDIjHPMzM6teUWXUl\n7SZpNvnOfZJ2kPQf/YzRzMwGgVqu8/gBsD/wIkBE3M/ySRLNzGw1VEvyICLmlRUtqeVxki6QtEDS\nAxW2fU3SUkkbFMrOkTRX0kxJOxbKj5D0qKRHJB1eS909PTBjRvptZmaNVUvymCdpdyAkDZf0z0Ct\n99O4iNRqWYGkMcC+wJOFsgOBrSJiG+AY4PxcPgo4GdgF2BWYImlkb5X29MCee8KkSem3E4iZWWPV\nkjw+R5occVNgPrAjcGwtB4+I3wMvV9h0FvD1srLJwCX5cXcAIyWNJiWfaRHxSkQsJN0U6oDe6p01\nCx56CBYvhtmz07KZmTVOLclj24j4ZESMjoiNIuJTwHb9rVDSIcC8iHiwbNOmQLF77OlcVl4+P5dV\nNWECjB8Pw4fD9tunZTMza5xarvP4IbBzDWV9krQWcBKpy6rP3es9fsmIEXDbbanFMX58Wjczs8ap\nmjwk7QbsDrxN0lcLm9YDhvazvq2AccD9kgSMAe6V9D5Si2Kzwr5jctl80n3Qi+XTq1UwderUZctd\nXV2MGNFVbVczs9VSd3c33d3dAzpG1YsEJe1F+tD+HHnwOusBrs83heq7Amlc3v9dFbb9Cdg5Il6W\ndBDwhYg4WNJE4AcRMTEPmN9NaukMycvvyeMf5cfzRYJmZnVq6D3MI+J3pHt5/DQinswVDAHWjYhF\nNQZ0GSkBbSjpKWBKRFxUrIbcPRURN0g6SNJjpDm0jsrlL0v6NilpBPCtSonDzMxap5b7eVxGan0s\nAe4idVudHRHfa3549XHLw8ysfk2ZngTYPrc0Pgz8GtgC+HQ/4jMzs0GiluQxXNJwUvK4LiLexPcw\nNzNbrdWSPH4EPAGsA9wqaSxQ05iHmZkNTjVPyb7Cg6RhEbG4CfEMiMc8zMzq19CzrQoHPbnKpv9b\nT0VmZjZ41HKF+WuF5bcAf0ftEyOamdkgVHe3laQ1gRsjoqspEQ2Au63MzOrXrFN1y61NmiLEzMxW\nU7WMeTzI8lNzhwJvw+MdZmartVquMB9bWF0MLOjEM63A3VZmZv3Rn26r3iZG3KDihiwiXqqnolZw\n8jAzq1+jT9W9h8LEhWUC2LKeiszMbPDo10WCncotDzOz+jXlbCtJfy9pZGF9fUkfrjGgCyQtkPRA\noex0SQ9Lminp55LWK2z7hqS5eft+hfIDJM2R9KikE2p/emZm1gy1DJjPjIgdy8rui4id+jy49H7g\nVeCSiHh3LvsgcEtELJV0KhAR8Q1J2wM/A3YhnQp8E7ANqdvsUWAf4BnStPCHRsScCvW55WFmVqdm\nXedRaZ9arkwnIn4PvFxWdlNELM2rt7P8mpFDgCsiYnFEPAHMBd6Xf+ZGxJN5Rt8rgMm11G9mZs1R\nS/K4W9KZkrbKP2eSBtMb4Z+AG/LypsC8wrb5uay8/OlcZmZmbVJL8vgS8AZwJelb/9+ALwy0Ykn/\nCrwZEZcP9FhmZtZafXY/RcRrwImNrFTSkcBBwN6F4vnAZoX1MblMwOYVyiuaOnXqsuWuri66uroG\nGq6Z2aDS3d1Nd3f3gI7R9FN1JY0Dro+Id+X1A4AzgEkR8WJhv9KA+a6kbqnfkgbMhwCPkAbMnwXu\nBA6LiJVm9vWAuZlZ/ZpyP4+BkHQZ0AVsKOkpYApwErAG8FtJALdHxLERMVvSVcBs4E3g2JwJlkj6\nIjCNlEguqJQ4zMysdVaLiwR7emDWLJgwAUaMaENgZmYdrFkXCZ4uaT1JwyXdLOnPkj7V/zBbq6cH\n9twTJk1Kv3t62h2Rmdmqr5azrfaLiEWkOwg+AWwNfL2ZQTXSrFnw0EOweDHMnp2WzcxsYGpJHsPz\n74OBqyPilSbG03ATJsD48TB8OGy/fVo2M7OBqWXA/DpJc4C/Ap+X9DbStR6rhBEj4LbbUotj/HiP\neZiZNUKvA+aShgATgTnAKxGxRNI6wIiIeK5FMdbMp+qamdWvoTeDKhy0pkkQO4GTh5lZ/Zo1MeLN\nkj6qfFGGmZlZLS2PHmAdYAlp3EOkadTX6/WBbeCWh5lZ/ZpyhXlEeIjZzMxWUNP0JJIOASbl1e6I\n+FXzQjIzs05XS7fVqaS7+/0sFx0G3B0R32hybHVzt5WZWf2adbbVA8COpbv/SRoK3Fe6rWwncfIw\nM6tfs862Ali/sDyyngrMzGzwqWXM4xTgPknTSWdaTQI6rsvKzMxap8+WR75N7ETgF8DPgd0i4opa\nDi7pAkkLctdXqWyUpGmSHpF0o6SRhW3nSJoraaakHQvlR0h6ND/m8HqeoJmZNV4tU7LfHBHPRsR1\n+ec5STfXePyLgP3Lyk4EboqIbYFbyK0YSQcCW0XENsAxwPm5fBRwMmnQfldgSjHhmJlZ61VNHpLe\nImkD4K25tbBB/hlHuk1snyLi98DLZcWTgYvz8sV5vVR+SX7cHcBISaNJyWdaRLwSEQtJdxQ8oJb6\nzcysOXob8zgG+AqwCXBvoXwRcO4A6twoIhYA5FbM6Fy+KTCvsN/Tuay8fD41Ji8zM2uOqskjIs4G\nzpb0pYj4YRNjqHZubb/m0po6deqy5a6uLrq6ulbY7lvSmtnqrru7m+7u7gEdo+p1HpL2johbJH2k\n0vaI+EVNFUhjgetL14VIehjoiogFkjYGpkfEdpLOz8tX5v3mAHsBH8j7fy6Xr7BfWV29XudRuiVt\n6d4et93mBGJm1ujrPPbKvz9U4efv6omLFVsR1wFH5uUjgV8Wyg8HkDQRWJi7t24E9pU0Mg+e75vL\n6uZb0pqZNUafV5gP6ODSZUAXsCGwAJgCXAtcDWwGPAl8PA+EI+lc0mD4a8BREXFvLj8S+FdSF9e/\nR8QlVeqrqeUxe3a6Ja1bHmZmDZ6eRNJXe3tgRJxZT0WtUMv0JD09viWtmVlRo6dkL320bku6xuK6\nvP4h4M76w+sMI0bAxIntjsLMbNVWy8SItwIHR0RPXh8B/E9ETOr1gW3giRHNzOrXrIkRRwNvFNbf\nyGVmZraaqmVixEuAOyX9d17/MMuvEDczs9VQTWdbSdoZ2DOv3hoR9zU1qn5yt5WZWf2aeT+PtYFF\n+arzpyVtUXd0ZmY2aNQyYD4FeC+wbUS8Q9ImwNURsUcrAqyHWx5mZvVrVsvj74FDSBfuERHPsPw0\nXjMzWw3VkjzeyF/nA0DSOs0NyczMOl0tyeMqST8C1pd0NHAT8F/NDcvMzDpZrWdb7QvsR5rg8MaI\n+G2zA+sPj3mYmdWvoXNbrYqcPMzM6tfMU3XNzMyWaVvykHS8pFmSHpD0M0lrSBon6XZJj0q6XNKw\nvO8akq6QNFfSDEmbtytuMzPrJXlIujn/Pq3RleZrRb4E7JzvMDgMOAw4DTgjIt4BLAQ+kx/yGeCl\niNgG+AFweqNjMjOz2vXW8ni7pN2BQyTtJGnn4k8D6h4KrJNbF2sBz5BuOfvzvP1i0jxaAJNZPp/W\nNcA+DajfzMz6qbeJEU8GvgmMAcpv/BTA3v2tNCKekXQG8BTwF2AacC/p1rNL825PA5vm5U2Befmx\nSyQtlLRBRLzU3xhKenrS7WknTPDNoczMalU1eUTENcA1kr4ZEd9uZKWS1ie1JsYCr5BuS3tAPYdo\nRByl29KW7izo29KamdWmzynZI+Lbkg4BSjd/6o6IXw2w3g8Cj5daDnm69z1IFyIOya2PMcD8vP98\n0j3Pn5E0FFivWqtj6tSpy5a7urro6uqqGsSsWSlxLF6c7mv+0EO+y6CZDX7d3d10d3cP6Bi1TIx4\nCvA+4Ge56DDgrog4qd+VSu8DLiDd3vZ14CLgLlKC+kVEXCnpP4H7I+J8SccCEyLiWEmHAh+OiEMr\nHLeu6zxKLY/Zs2H77d3yMLPVU1MuEpT0ALBjaSwif/O/L58l1W95tt5DgTeB+4D/Q2ptXAGMymWf\niog3Ja0JXArsBLwIHBoRT1Q4Zt0XCfb0LO+2cuIws9VRM5NHV6GLaQNS19WAkkcz+ApzM7P69Sd5\n1HIb2lOA+yRNJw1UTwJO7Ed8ZmY2SNQ6MeLbSeMTAHdGxHNNjaqf3PIwM6ufJ0Z08jAzq5snRjQz\ns5Zw8jAzs7r1mjwkDZU0p1XBtFtPD8yYkX6bmVl1vSaPiFgCPLI6TIFeumBw0qT02wnEzKy6Wk7V\nHQU8JOlO4LVSYUQc0rSo2sBTlZiZ1a6W5PHNpkfRASZMSFeZl6YqGT++3RGZmXWuWq/zGAtsExE3\nSVobGBoRHdexM9BTdT1ViZmtjpo1PcnRwGeBDSJiK0nbAOdHRMfdkMnXeZiZ1a9Z13l8gTRd+iKA\niJgLbFR/eGZmNljUkjxej4g3Siv5trH+em9mthqrJXn8TtJJwFqS9iXd9e/6gVYsaaSkqyU9LOkh\nSbtKGiVpmqRHJN0oaWRh/3MkzZU0U9KOA63fzMz6r5bkcSLwZ+BB4BjgBuDfGlD32cANEbEdsAMw\nJ9d1U0RsC9wCfANA0oHAVhGxTY7h/AbUb2Zm/VTr2VZrAO8kdVc9UuzG6lel0nqkG0ptVVY+B9gr\nIhZI2hiYHhHbSTo/L1+Z93uYdI+RBWWPb+iAeU9Puv5jwgSffWVmg1dTBswlHQz8ETgHOBd4LLcE\nBmIL4AVJF0m6V9KP8ynAo0sJIU/7Pjrvvykwr/D4+bmsaXzFuZlZdbV0W50BfCAiuiJiL+ADwFkD\nrHcYsDNwXkTsTLpy/URWHohv28B8pSvOzcwsqeUK856IeKyw/jgw0O/hTwPzIuLuvP5zUvJYIGl0\nodvq+bx9PrBZ4fFjctlKpk6dumy5q6uLrq6ufgXoK87NbLDq7u6mu7t7QMeoOuYh6SN5cV9gLHAV\nqSXwMeBIl04oAAAQKUlEQVSpiDh2QBVLvwOOjohHJU0B1s6bXoqI0ySdCKwfESdKOgj4QkQcLGki\n8IOIWGnmqWaMefiKczMb7Bp6hbmki3p7YEQcVU9FFY6/A/ATYDipNXMUMJSUpDYDngQ+HhEL8/7n\nAgeQuriOioh7KxzTV5ibmdXJt6F18jAzq1t/kkefYx6StgC+BIwr7j/YpmQ3M7Pa1TJgfi1wAemq\n8qXNDcfMzFYFtSSPv0XEOU2PxMzMVhm1TMn+CWAbYBrweqm80oB1uzVzzMNXm5vZYNWUMQ/gXcCn\ngb1Z3m0VeX21ULravHTa7m23OYGY2eqtluTxMWDLgc5ntSrz/c3NzFZUy/Qks4D1mx1IJytdbT58\nuK82NzOD2sY8uoF3A3ex4phHx52q2+wxD19tbmaDUbPuYb5XpfKI+F09FbWCLxI0M6ufrzB38jAz\nq1uzrjDvYfnU6GuQ5qJ6LSLWqz9EMzMbDPpMHhGxrIdfkoDJgM81MjNbjdVyttUykVwL7N+keFYJ\nPT0wY4bvLmhmq69auq0+UlgdArwX+FvTIupwvmDQzKy2lseHCj/7k+4iOLkRlUsaku9hfl1eHyfp\ndkmPSrpc0rBcvoakKyTNlTRD0uaNqL8/fHtaM7PaxjwGdNOnPhwHzAZKg++nAWdExNWS/hP4DPCj\n/PuliNhG0j8CpwOHNjGuqnx7WjOz3u8keHIvj4uI+PaAKpbGABcB3wG+GhGHSPozMDoilubbzU6J\niAMl/SYv3yFpKPBcRLytwjFbcqquLxg0s8GkP6fq9tZt9VqFH0itgBP6FeGKzgK+Tj4NWNKGwMsR\nUZp88Wlg07y8KTAPICKWAAslbdCAGPplxIg0t5UTh5mtrqp2W0XEGaVlSSNIXUxHAVcAZ1R7XC0k\nHQwsiIiZkrqKm2o9RLUNU6dOXbbc1dVFV1dXtV0bxtO1m9mqpLu7m+7u7gEdo9crzPO3+68CnwQu\nBs6OiJcHVGM67neBTwGLgbWAEaQ7Fu4HbFxDt9WzEbFRheO2/Apzn31lZqu6hnZbSfoeaTLEHuBd\nETG1EYkDICJOiojNI2JL0sD3LRHxKWA6aQp4gCOAX+bl6/I6efstjYijEXz2lZmtjnobMF9KmkV3\nMcunJ4HUZRSNmp4kT7z4tTxgvgWpW2wUcB/wqYh4U9KawKXATsCLwKER8USFY7Wt5VE6+8otDzNb\n1XhixDZNjOizr8xsVebk0QGz6nrw3MxWNY0+VdfqVOrCmjQp/fbcV2Y2WDl5NJAHz81sdeHk0UDl\n9zrffHPPvmtmg5PHPBqsNHi++eZw0EG+/sPMOp/HPDpAaeqSJ590F5aZDV5OHk1S3oU1frxvImVm\ng4e7rZqoeP0HeBoTM+tMvs6jw5JH0YwZ6RTexYtTa+TWW1P3lplZu3nMo4P5TCwzG0zc8mghn4ll\nZp3ILY8O5zOxzGywcPJoA3dhmdmqri3JQ9IYSbdIekjSg5K+nMtHSZom6RFJN0oaWXjMOZLmSpop\nacd2xN0oI0akrqpbb4UbbkhdWMX5sHxKr5l1una1PBYDX42I8cBuwBckvRM4EbgpIrYl3fDpGwCS\nDgS2iohtgGOA89sTduNU68K6805Prmhmna8tySMinouImXn5VeBhYAwwmXS7W/LvyXl5MnBJ3v8O\nYKSk0S0NuknKu7AiVk4mboWYWacZ1u4AJI0DdgRuB0ZHxAJICaaQIDYF5hUeNj+XLWhdpM1R6sIq\nXkw4fnxKHNtuC1/5CsyZ47OyzKyztDV5SFoXuAY4LiJelVR+nm3d591OnTp12XJXVxddXV0DCbEl\nSl1YJaVk8uqrcOCBK7ZC1l7bN5oys4Hp7u6mu7t7QMdo23UekoYBvwJ+HRFn57KHga6IWCBpY2B6\nRGwn6fy8fGXebw6wV6mVUjhmR1/nUa/i/dG33TaVFVsh4LsWmtnArWrXeVwIzC4ljuw64Mi8fCTw\ny0L54QCSJgILyxPHYFQ8K+uss1LiqDaw/swzHhsxs9ZpS8tD0h7ArcCDpK6pAE4C7gSuAjYDngQ+\nHhEL82POBQ4AXgOOioh7Kxx3ULU8ioqtkO23h+9/f3mX1rBhMG4cPPFEapXccEM6i8stEjOrhSdG\nHMTJAyrP0jt7NowdmxJHpUTi7i0z64uTxyBPHuXK58oqTyTDh8Ovfw1f+9rypONWiZmVc/JYzZJH\nUaVE4u4tM6uFk8dqnDyKGtG9VW3ZScZs8HHycPKoqN7urXe+Mz1uzpwVlz2GYjY4OXk4efSplu6t\noUNBWnnZYyhmg5OTh5NHXap1b5UuSHzkkRWX+xpD6a3ry4nFrHM5eTh5DEh5Mqm2XGkMpbeuL7dQ\nzDqbk4eTR0vU2/U1kMH5nh4P4ps1m5OHk0fL1dL11d/B+dKNsuodxK+27CRjVpmTh5NH21Xq+urv\n4Px558Gxx9Y3iO8kY1Y/Jw8nj47Vn8H5UsujnkH8VieZat1q9XS5Ffdz4rJ26E/yICIGzU96OrYq\nWLQoYsaM9Lvacq37LVoUscMOEcOHR0yYkH7Kl3fYIeK3v40YNiwCIoYOrbw8fHjETTel/YcNW36M\n8uUddoiYP7/yfr1t622/0vr//u/y51VaLr0WlbZ1+n7W+fJnZ32ft/U+oJN/nDxWX+1IMj/+cf3b\nettv2LCIrbduXDLqhP3KE2Lpb+WE1lkGffIgTck+B3gUOKHC9sa8kjZoNTLJlD4w69nW236lxNGo\nZNQJ+xUTYjNaa81OaKtCq64RdQ3q5EG6cdVjwFhgODATeGfZPvV8jjTN9OnT2x1CR8QQ0Rlx9CeG\ngXarVdp23nnTe92v0cmo0n5Dh05v6PH62q+YEFdMMtPbntCGDJne0qRVbb8hQ6a3tHu00n6DPXlM\nJN2ytrR+Ynnro1OSx5QpU9odQkfEENEZcXRCDBG1xdGfRFXPfp/5zJSGHq+v/YpJZcUPzClNS5C1\nJ7QpHdJam9LS7tHK+9WfPIb1b2y+LTYF5hXWnwbe16ZYzJpixAiYOHH5enG5t2217jdmzPIzuhpx\nvFr2u+225WfalW6tfPzx6dbKsPK2vmY5GOh+pVPHH3wwna138MFpe19n9TVrvwcfhO22q/94jYxp\n1izqtiolDzNbBZUnlREjVkxijU6QtSa0UgJrVdKqtl+lRFrr8RoV03rrUbdV5joPSROBqRFxQF4/\nkdTUOq2wz6rxZMzMOkwM1osEJQ0FHgH2AZ4F7gQOi4iH2xqYmdlqaJXptoqIJZK+CEwjnXl1gROH\nmVl7rDItDzMz6xxD2h1Ao0g6QNIcSY9KOqEN9Y+RdIukhyQ9KOnLrY6hLJ4hku6VdF2b6h8p6WpJ\nD+fXZNc2xXG8pFmSHpD0M0lrtKjeCyQtkPRAoWyUpGmSHpF0o6SRbYjh9Pw3mSnp55L6MVQ68DgK\n274maamkDdoRg6Qv5dfjQUmnNjOGanFI2kHSDEn3SbpT0nubHEPFz6q635/1ntvbiT/UcAFhC2LY\nGNgxL69LGp9paQxl8RwP/D/gujbV/1PgqLw8DFivDTFsAjwOrJHXrwQOb1Hd7wd2BB4olJ0G/Ete\nPgE4tQ0xfBAYkpdPBU5px2uRy8cAvwH+BGzQhteii9QNPiyvv7VN74sbgf3y8oHA9CbHUPGzqt73\n52BpebwPmBsRT0bEm8AVwORWBhARz0XEzLz8KvAw6dqUlpM0BjgI+Emb6l8P2DMiLgKIiMURsagd\nsQBDgXUkDQPWBp5pRaUR8Xvg5bLiycDFefli4MOtjiEiboqIpXn1dtIHeFNVeS0AzgK+3uz6e4nh\n86QPyMV5nxfaFMdSoPQtf31gfpNjqPRZNYY635+DJXlUuoCwLR/cAJLGkb5d3NGmEEr/lO0a0NoC\neEHSRbnr7MeS1mp1EBHxDHAG8BTpH3JhRNzU6jgKNoqIBTm254CN2hgLwD8Bv25HxZIOAeZFxIPt\nqD97BzBJ0u2Spje7u6gXxwPfl/QUcDrwjVZVXPisuh0YXc/7c7Akj44haV3gGuC4nNVbXf/BwIL8\nzUL5p9WGATsD50XEzsBfSNPJtJSk9UnfpsaSurDWlfSJVsfRi7adrSLpX4E3I+KyNtS9FnASMKVY\n3Oo4SO/TURExEfgX4Ko2xACpBXRcRGxOSiQXtqLSCp9V5e/HXt+fgyV5zAc2L6yPoclNv0py18g1\nwKUR8ctW15/tARwi6XHgcuADki5pcQxPk75V3p3XryElk1b7IPB4RLwUEUuAXwC7tyGOkgWSRgNI\n2hh4vh1BSDqS1K3ZrkS6FTAOuF/Sn0j/r/dIanVLbB7pPUFE3AUslbRhi2MAOCIirs1xXEMLpl2q\n8llV1/tzsCSPu4CtJY3NZ9McCrTjLKMLgdkRcXYb6gYgIk6KiM0jYkvS63BLRBze4hgWAPMkvSMX\n7QPMbmUM2VPARElvkaQcRyuvDSpv+V0HHJmXjwBa8QVjhRgkHUDq0jwkIl5vQf0rxRERsyJi44jY\nMiK2IH3Z2Ckimp1My/8e1wJ7A+T36vCIeLHJMVSKY76kvXIc+5BuOdFslT6r6nt/Nvvsglb9kO71\n8QgwFzixDfXvASwhnel1H3AvcECbX5O9aN/ZVjuQkvpM0re7kW2KYwopYTxAGgQc3qJ6LyMNzr9O\nSmJHAaOAm/L7dBqwfhtimAs8md+f9wL/0Y7Xomz74zT/bKtKr8Uw4FLgQeBuYK82vS92z/XfB8wg\nJdJmxlDxswrYoJ73py8SNDOzug2WbiszM2shJw8zM6ubk4eZmdXNycPMzOrm5GFmZnVz8jAzs7o5\neZgBkjbIU2LfK+lZSU/n5ZclzWpCfXtJur7Ox0yXtNKV+pKOkPTDxkVn1rdV5k6CZs0UES8BOwFI\nOhl4NSLOlDQW6PNDXtLQSFOg1FVt/ZG25FhmfXLLw2xl5ZP0DcszA8+S9BtJa8KylsBZku4Cvizp\nrZKukXRH/tkt77dXoVVzj6R18nFHaPkNsy5dVrm0T973fkk/kTR8pQClo/JNe24nXTFs1lJOHmZ9\n2wb4YURMAF4BPlrYNjwidomIs4CzgTMjYlfgH4AL8j5fA46NNMPwnsBfc/mOwJeB7YGtJO2eE9NF\nwMciYgfSzc0+XwwmT1o3FdiNdHOh7Rv8fM365ORh1rfHY/l9J+4hzQhbcmVh+YPAuZLuI00yt66k\ntYE/AGdJ+hJpCvDSzZjujIhnI80RNDMfd9tc3x/zPhcDk8ri2ZV0t7mXIt3I6ErMWsxjHmZ9K84+\nuwR4S2H9tcKygF0j3c2y6DRJvwIOBv4gab8qxy39P9Zyb4t23P/CbBm3PMz6VusH9TTguGUPknbI\nv7eMiIci4nTSTMPv7OUYjwBjJW2Z1z8NdJftcwfpDnij8njIx2qMz6xhnDzM+lbtTKby8uOA9+aB\n7lnAMbn8K5IelHQ/8AaVb/0aAJHusXEUcE3efwnwo7J9niONedwO3EZ77pViqzlPyW5mZnVzy8PM\nzOrm5GFmZnVz8jAzs7o5eZiZWd2cPMzMrG5OHmZmVjcnDzMzq5uTh5mZ1e3/AycrGZASWvFqAAAA\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f40384e8cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "# Talk editors in at least one cluster by threshold\n",
    "\n",
    "plt.plot(total_clustered[0], total_clustered[1], '.')\n",
    "plt.title(\"Talk editors in at least one cluster by threshold\")\n",
    "plt.xlabel('Threshold')\n",
    "plt.xticks(range(0, 51, 2))\n",
    "plt.axis([0, 20, 0, 2000])\n",
    "plt.yticks(range(0, 2001, 200))\n",
    "plt.ylabel('Number of editors clustered')\n",
    "plt.savefig('/home/michael/school/research/wp/pca_editors_classified.png', dpi=400)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "47.408197826\n",
      "79.256529847\n",
      "68.905077319\n",
      "92.274107605\n",
      "48.831714948\n",
      "20.725912078\n",
      "41.368015891\n",
      "35.723681858\n",
      "67.446532604\n",
      "58.238153128\n"
     ]
    }
   ],
   "source": [
    "# Max loadings in each column\n",
    "for i in range(1,11):\n",
    "    print(max(factors_talk[\"Factor{:d}\".format(i)]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "2\n",
      "3\n",
      "4\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# TALK PAGE DATA\n",
    "# Gather talk page text from editors\n",
    "talk_csvpath = '/home/michael/school/research/wp/wikipedia/data/talk/ipc_utf8_talkpages.csv'\n",
    "talk_data = pd.read_csv(talk_csvpath)\n",
    "talk_txt = {} # talk_txt[editor] = all_text\n",
    "for ed in list(talk_eds):\n",
    "    #for _, row in talk_data[talk_data['username']==ed].iterrows():\n",
    "    #    debug_here()\n",
    "    #    print(row['post_text'])\n",
    "    talk_txt[ed] = \" \".join([str(row['post_text']) for _, row in talk_data[talk_data['username']==ed].iterrows()])\n",
    "\n",
    "# print(len(talk_txt))\n",
    "\n",
    "# Build dict with top and bottom quantile of each factor\n",
    "factor_eds = {i: {} for i in range(1,11)}\n",
    "div = 100/edpercent\n",
    "for i in range(1, 11):\n",
    "    factor_eds[i]['top'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=False)[:round(len(factors_talk)/div)]\n",
    "    factor_eds[i]['bottom'] = factors_talk['Factor{:d}'.format(i)].sort_values(ascending=True)[:round(len(factors_talk)/div)]\n",
    "\n",
    "# Build up factor training text from top quantiles\n",
    "factor_txt = {}\n",
    "for i in range(1, 5):\n",
    "    #for ed in factor_eds[i]['top'].index:\n",
    "    #    if str(ed) not in talk_txt.keys():\n",
    "    #        print(ed)\n",
    "    factor_txt[i] = ' '.join([talk_txt[ed] for ed in factor_eds[i]['top'].index])\n",
    "\n",
    "# Clean up text\n",
    "numberref = \"(?:[0-9](?:,|\\.)?)+(?:x|k|m|st|nd|rd|th)?\"\n",
    "numberspace = re.compile(\"{0} \".format(numberref))\n",
    "spacenumber = re.compile(\" {0}\".format(numberref))\n",
    "underscores = re.compile(\"_\")\n",
    "\n",
    "# two options for datestring: \n",
    "#    xx:xx . yyy # . # . UTC .\n",
    "#    xx:xx . # yyy # . UTC .\n",
    "timeref = \"[0-2]?[0-9]:[0-9][0-9]\"\n",
    "months = [  \"January\", \"February\",    \"March\",    \"April\",\n",
    "                \"May\",     \"June\",     \"July\",   \"August\",\n",
    "          \"September\",  \"October\", \"November\", \"December\"]\n",
    "def getRef(month):\n",
    "     if len(month)>3:\n",
    "        return \"(?:{0}(?:\\.|{1})?)\".format(month[:3],month[3:])\n",
    "     else:\n",
    "        return \"(?:{0})\".format(month)\n",
    "    \n",
    "# getRef = lambda(month):\"(?:{0}(?:\\.|{1})?)\".format(month[:3],month[3:]) if len(month)>3 \\\n",
    "#                   else \"(?:{0})\".format(month)\n",
    "\n",
    "monthref = \"(?:{0})\".format(\"|\".join([getRef(month) for month in months]))\n",
    "UTCref = \"(?:(?:(?:\\(UTC\\))|(?:\\(UTC ?(?:\\+|-)[0-1]?[0-9]\\))|(?:UTC(?: ?(?:\\+|-)[0-1]?[0-9]\\)?)?)))\"\n",
    "tsConfigs = [\"(?:{0},? \\. {1} {3} \\. {3} \\. {2} \\.)\",\n",
    "             \"(?:{0},? \\. {3} {1} {3} \\. {2} \\.)\",\n",
    "             \"(?:{3} {1} {3} {0} {2})\",\n",
    "             \"(?:{0},? {3} {1} {3} {2})\"]\n",
    "timestamp = re.compile(\"(?:{0})\".format(\"|\".join([config.format(timeref, monthref, UTCref, numberref) \n",
    "                                              for config in tsConfigs])))\n",
    "time = re.compile(\" {0} \".format(timeref))\n",
    "IPnumref = \"(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\"\n",
    "IPaddress = re.compile(\"(?:{0}\\.){{3}}{0}\".format(IPnumref))\n",
    "# name = re.compile(\"|\".join(list(talk_eds)))\n",
    "\n",
    "maxlen = 0\n",
    "for key in factor_txt:\n",
    "    segment = factor_txt[key]\n",
    "    segment = timestamp.sub(\"TIMESTAMP\",segment)\n",
    "    segment = IPaddress.sub(\"IPADDRESS\",segment)\n",
    "    segment = time.sub(\" TIME \",segment)\n",
    "    segment = spacenumber.sub(\" #\",segment)\n",
    "    segment = numberspace.sub(\"# \",segment)\n",
    "    segment = underscores.sub(\" \",segment)\n",
    "#     segment = name.sub(\"NAME\", segment)\n",
    "    factor_txt[key] = segment\n",
    "    \n",
    "for i in range(1,5):\n",
    "    print(i)\n",
    "    factor_txt[i] = factor_txt[i].replace(':', '')\n",
    "    for ed in talk_eds:\n",
    "        factor_txt[i] = factor_txt[i].replace('--{0}'.format(ed), '--NAME')\n",
    "        factor_txt[i] = factor_txt[i].replace('-{0}'.format(ed), '-NAME')\n",
    "        factor_txt[i] = factor_txt[i].replace('@{0}'.format(ed), '-NAME')\n",
    "        factor_txt[i] = factor_txt[i].replace(' {0} '.format(ed), ' NAME ')\n",
    "        factor_txt[i] = factor_txt[i].replace('NAME (talk)'.format(ed), 'NAME')\n",
    "    \n",
    "len(factor_txt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Write out factor texts\n",
    "data = pd.DataFrame(factor_txt, index=['text']).transpose()\n",
    "data['factor'] = data.index\n",
    "data = data[data.columns.tolist()[-1:] + data.columns.tolist()[:-1]]\n",
    "\n",
    "data.to_csv('/home/michael/school/research/wp/wikipedia/data/top{:d}percent_editortext.csv'.format(edpercent), index=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "Column 1\n",
      "Jalapenos do exist    47.408198\n",
      "Tiamut                21.014307\n",
      "Nishidani             20.313740\n",
      "The Squicks           16.979149\n",
      "Lihaas                15.702478\n",
      "Okedem                15.490496\n",
      "Night w               14.616990\n",
      "Al Ameer son          14.043197\n",
      "Factomancer           13.642668\n",
      "Tewfik                12.987740\n",
      "Name: Factor1, dtype: float64\n",
      "Column 1\n",
      "WLRoss                  4.002824\n",
      "Jethro B                3.979675\n",
      "JiHymas@himivest.com    3.848054\n",
      "Olegwiki                3.841054\n",
      "No More Mr Nice Guy     3.762094\n",
      "6SJ7                    3.742516\n",
      "Pedrito                 3.722863\n",
      "Tpbradbury              3.615314\n",
      "Fences and windows      3.588243\n",
      "John Bahrain            3.551108\n",
      "Name: Factor1, dtype: float64\n",
      "\n",
      "2\n",
      "Column 1\n",
      "Reenem                        79.256530\n",
      "Jalapenos do exist            25.125361\n",
      "TheCuriousGnome               23.719465\n",
      "Gilabrand                     22.924933\n",
      "Marokwitz                     22.496192\n",
      "Vanished user lt94ma34le12    19.553602\n",
      "Sceptic Ashdod                17.371216\n",
      "Sean.hoyland                  17.296738\n",
      "Ynhockey                      16.382318\n",
      "Nableezy                      14.482380\n",
      "Name: Factor2, dtype: float64\n",
      "Column 1\n",
      "Trahelliven                          2.871526\n",
      "Tazmaniacs                           2.785037\n",
      "Yambaram                             2.751561\n",
      "PluniAlmoni                          2.737448\n",
      "Mhym                                 2.659725\n",
      "Gimmetrow                            2.648123\n",
      "Ohnoitsjamie                         2.615076\n",
      "Magioladitis                         2.614082\n",
      "Vanished user skj3ioo3jwifjsek35y    2.569426\n",
      "Flymeoutofhere                       2.561749\n",
      "Name: Factor2, dtype: float64\n",
      "\n",
      "3\n",
      "Column 1\n",
      "Wickey-nl           68.905077\n",
      "Ashley kennedy3     44.071626\n",
      "Al Ameer son        41.513585\n",
      "Jayjg               21.047769\n",
      "Wikifan12345        13.929464\n",
      "Harlan wilkerson    12.589968\n",
      "Gilabrand           11.240105\n",
      "Night w             10.989259\n",
      "GHcool               9.937438\n",
      "Hertz1888            9.502926\n",
      "Name: Factor3, dtype: float64\n",
      "Column 1\n",
      "Emmette Hernandez Coleman    2.739489\n",
      "Andyvphil                    2.714086\n",
      "ElComandanteChe              2.573571\n",
      "MPerel                       2.558370\n",
      "Jmabel                       2.531181\n",
      "Carolmooredc                 2.430793\n",
      "Jhoman                       2.395869\n",
      "Khoikhoi                     2.371244\n",
      "Statesman1                   2.360858\n",
      "Koavf                        2.337126\n",
      "Name: Factor3, dtype: float64\n",
      "\n",
      "4\n",
      "Column 1\n",
      "Huldra             92.274108\n",
      "Al Ameer son       48.939697\n",
      "Ashley kennedy3    26.100198\n",
      "Tiamut             24.109692\n",
      "Nishidani          11.550845\n",
      "Padres Hana         8.412404\n",
      "Reenem              6.878799\n",
      "Ynhockey            6.679978\n",
      "Ian Pitchford       6.663346\n",
      "Ykantor             6.206679\n",
      "Name: Factor4, dtype: float64\n",
      "Column 1\n",
      "Danlaycock        1.150490\n",
      "Noon              1.094017\n",
      "Paraphelion       1.076257\n",
      "BD2412            1.059132\n",
      "Marek69           1.040581\n",
      "Fjmustak          1.036534\n",
      "Jim Fitzgerald    1.036465\n",
      "Alaqsa81090       1.035950\n",
      "Talknic           1.030973\n",
      "FunkMonk          1.030063\n",
      "Name: Factor4, dtype: float64\n",
      "\n",
      "5\n",
      "Column 1\n",
      "Marokwitz               48.831715\n",
      "Sa.vakilian             35.621876\n",
      "JiHymas@himivest.com    14.925126\n",
      "Jayjg                    9.648516\n",
      "Organ123                 8.423445\n",
      "G-Dett                   8.201685\n",
      "Egeymi                   7.545191\n",
      "TheCuriousGnome          7.203205\n",
      "Wikifan12345             7.064997\n",
      "WLRoss                   6.343017\n",
      "Name: Factor5, dtype: float64\n",
      "Column 1\n",
      "Middayexpress     2.623897\n",
      "Dmyersturnbull    2.574149\n",
      "Ahangar-e-Gaz     2.573454\n",
      "Magioladitis      2.572393\n",
      "IZAK              2.531188\n",
      "Wickey-nl         2.528443\n",
      "Gabe2000000       2.520036\n",
      "Brewcrewer        2.485352\n",
      "Sean.hoyland      2.380052\n",
      "386-DX            2.375829\n",
      "Name: Factor5, dtype: float64\n",
      "\n",
      "6\n",
      "Column 1\n",
      "Michael Safyan    20.725912\n",
      "Carolmooredc      20.374844\n",
      "Wikifan12345      17.382228\n",
      "NYScholar         12.587437\n",
      "Kendrick7         11.551243\n",
      "Nableezy          10.692401\n",
      "Sm8900            10.656421\n",
      "Huldra             8.976700\n",
      "Atrix20            8.647944\n",
      "Sean.hoyland       8.257613\n",
      "Name: Factor6, dtype: float64\n",
      "Column 1\n",
      "Boodlesthecat       2.929852\n",
      "Grenavitar          2.888791\n",
      "Tazmaniacs          2.751421\n",
      "NoCal100            2.700934\n",
      "Palmiro             2.639004\n",
      "Shuki               2.516163\n",
      "Jmabel              2.440812\n",
      "Soosim              2.437262\n",
      "IronDuke            2.424388\n",
      "Fayenatic london    2.423473\n",
      "Name: Factor6, dtype: float64\n",
      "\n",
      "7\n",
      "Column 1\n",
      "Shamir1          41.368016\n",
      "Gilabrand        32.840283\n",
      "Malik Shabazz    26.844886\n",
      "Al Ameer son     21.207587\n",
      "Nableezy         19.737863\n",
      "Sm8900           16.810542\n",
      "Wikifan12345     15.425726\n",
      "Ian Pitchford    14.454873\n",
      "Haberstr         14.201435\n",
      "Sa.vakilian      12.152741\n",
      "Name: Factor7, dtype: float64\n",
      "Column 1\n",
      "Soosim         3.060464\n",
      "Trahelliven    2.991244\n",
      "Flayer         2.976850\n",
      "CJCurrie       2.959572\n",
      "Yahel Guhan    2.928325\n",
      "NYCJosh        2.907845\n",
      "Tickle me      2.905143\n",
      "Tpbradbury     2.899487\n",
      "A Sniper       2.779240\n",
      "Yoninah        2.706085\n",
      "Name: Factor7, dtype: float64\n",
      "\n",
      "8\n",
      "Column 1\n",
      "Malik Shabazz         35.723682\n",
      "Tiamut                15.421296\n",
      "Nableezy              15.234778\n",
      "Sean.hoyland          14.673485\n",
      "Jalapenos do exist    13.448804\n",
      "Marokwitz             12.374105\n",
      "Gilabrand             11.737755\n",
      "SlimVirgin            11.347515\n",
      "Former user 2         10.723162\n",
      "GHcool                 9.895858\n",
      "Name: Factor8, dtype: float64\n",
      "Column 1\n",
      "Canadian Monkey    2.791087\n",
      "Hellomrkrinkle     2.785536\n",
      "Ceedjee            2.773030\n",
      "Dan Murphy         2.769187\n",
      "Samuel B52         2.729454\n",
      "WLRoss             2.721022\n",
      "Vice regent        2.695413\n",
      "Epeefleche         2.643058\n",
      "Beit Or            2.635687\n",
      "Joshdboz           2.578901\n",
      "Name: Factor8, dtype: float64\n",
      "\n",
      "9\n",
      "Column 1\n",
      "Former user 2    67.446533\n",
      "Armon            22.990481\n",
      "Tewfik           21.594469\n",
      "Jaakobou         21.040712\n",
      "Jayjg            16.187500\n",
      "Beit Or          15.953097\n",
      "Hertz1888        15.635665\n",
      "Haberstr         13.529184\n",
      "Sean.hoyland     13.281159\n",
      "Kasaalan         12.931234\n",
      "Name: Factor9, dtype: float64\n",
      "Column 1\n",
      "Mystichumwipe     2.285023\n",
      "Againme           2.272011\n",
      "Ling.Nut          2.262635\n",
      "Bibigon           2.251845\n",
      "Grace Note        2.224150\n",
      "Coroebus          2.221875\n",
      "Mschiffler        2.187996\n",
      "Hope&amp;Act3!    2.116364\n",
      "Rami R            2.111849\n",
      "Ethelh            2.055344\n",
      "Name: Factor9, dtype: float64\n",
      "\n",
      "10\n",
      "Column 1\n",
      "Nableezy          58.238153\n",
      "Jayjg             25.020719\n",
      "Sean.hoyland      16.287192\n",
      "CJCurrie          14.809520\n",
      "Nishidani         13.462572\n",
      "Michael Safyan    12.513016\n",
      "Urthogie          10.409443\n",
      "Factomancer       10.188248\n",
      "Zero0000           9.898999\n",
      "Wickey-nl          9.137819\n",
      "Name: Factor10, dtype: float64\n",
      "Column 1\n",
      "Burgas00      2.969112\n",
      "MPerel        2.931329\n",
      "Sleetman      2.925925\n",
      "Ronhjones     2.873208\n",
      "SlimVirgin    2.836675\n",
      "Oneworld25    2.763414\n",
      "DerHexer      2.754988\n",
      "AladdinSE     2.696401\n",
      "Alatari       2.692542\n",
      "Froy1100      2.623693\n",
      "Name: Factor10, dtype: float64\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Qualitatively look at top and bottom deciles\n",
    "for i in range(1, 11):\n",
    "    print(i)\n",
    "    print(factor_eds[i]['top'][:10])\n",
    "    print(factor_eds[i]['top'][50:60])\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "312\n",
      "510\n",
      "170\n"
     ]
    }
   ],
   "source": [
    "# Compare with WikiProject lists\n",
    "# Load in WikiProject lists\n",
    "proj_pal = set([name[0] for name in pd.read_csv('/home/michael/school/research/wp/wikipedia/data/wikiproject_palestine.txt', header=None).values.tolist()])\n",
    "print(len(proj_pal))\n",
    "proj_isr = set([name[0] for name in pd.read_csv('/home/michael/school/research/wp/wikipedia/data/wikiproject_israel.txt', header=None).values.tolist()])\n",
    "print(len(proj_isr))\n",
    "proj_collab = set([name[0] for name in pd.read_csv('/home/michael/school/research/wp/wikipedia/data/wikiproject_israel_palestine_collaboration.txt', header=None).values.tolist()])\n",
    "print(len(proj_collab))\n",
    "\n",
    "rows = [['pal_overlap', 'isr_overlap', 'collab_overlap']]\n",
    "for i in range(1, 11):\n",
    "    row = []\n",
    "    #print(i)\n",
    "    top = set(factor_eds[i]['top'].index.tolist())\n",
    "    overlap = top.intersection(proj_pal)\n",
    "    #print(\"\\tPal overlap: {:f}\".format(len(overlap)/len(dec)))\n",
    "    row.append(len(overlap)/len(dec))\n",
    "    overlap = top.intersection(proj_isr)\n",
    "    #print(\"\\tIsr overlap: {:f}\".format(len(overlap)/len(dec)))\n",
    "    row.append(len(overlap)/len(dec))\n",
    "    overlap = top.intersection(proj_collab)\n",
    "    #print(\"\\tCollab overlap: {:f}\".format(len(overlap)/len(dec)))\n",
    "    row.append(len(overlap)/len(dec))\n",
    "    rows.append(row)\n",
    "    #print()\n",
    "data = pd.DataFrame(rows[1:]).transpose()\n",
    "data"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}