ICC_17_11_2018.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\importlib\\_bootstrap.py:219: RuntimeWarning: cymem.cymem.Pool size changed, may indicate binary incompatibility. Expected 48 from C header, got 64 from PyObject\n",
      "  return f(*args, **kwds)\n",
      "C:\\ProgramData\\Anaconda3\\lib\\importlib\\_bootstrap.py:219: RuntimeWarning: cymem.cymem.Address size changed, may indicate binary incompatibility. Expected 24 from C header, got 40 from PyObject\n",
      "  return f(*args, **kwds)\n"
     ]
    }
   ],
   "source": [
    "import spacy\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from collections import Counter\n",
    "from glob import glob\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.decomposition import TruncatedSVD\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "# Display plots in this notebook, instead of externally. \n",
    "from pylab import rcParams\n",
    "rcParams['figure.figsize'] = 16, 8\n",
    "%matplotlib inline\n",
    "\n",
    "# The following are optional dependencies. \n",
    "# Feel free to comment these out. \n",
    "# Sent2tree uses the sent2tree.py module in this repository. \n",
    "from sent2tree import sentenceTree\n",
    "import ete3 \n",
    "import seaborn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\importlib\\_bootstrap.py:219: RuntimeWarning: cymem.cymem.Pool size changed, may indicate binary incompatibility. Expected 48 from C header, got 64 from PyObject\n",
      "  return f(*args, **kwds)\n",
      "C:\\ProgramData\\Anaconda3\\lib\\importlib\\_bootstrap.py:219: RuntimeWarning: cymem.cymem.Address size changed, may indicate binary incompatibility. Expected 24 from C header, got 40 from PyObject\n",
      "  return f(*args, **kwds)\n"
     ]
    }
   ],
   "source": [
    "# This command might take a little while. \n",
    "#nlp = spacy.load('en')\n",
    "nlp = spacy.load('en_core_web_lg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "pdf_text_refined = open('witness_text_clean.txt').read()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parse the texts. These commands might take a little while. \n",
    "pdf_text = nlp(pdf_text_refined)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "My"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pdf_text[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Exploring Named Entities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'CARDINAL', 'DATE', 'GPE', 'ORG', 'PERSON', 'TIME'}"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set([w.label_ for w in pdf_text.ents])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[Goma, Goma, Congo]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[ent for ent in pdf_text.ents if ent.label_ == 'GPE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Marcus BRODY'}"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set(list([ent.string.strip() for ent in pdf_text.ents if ent.label_ == 'PERSON']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ngrams(doc, n): \n",
    "    doc = [word for word in doc \n",
    "           if word.is_alpha # Get rid of punctuation\n",
    "           if not word.string.isupper()] # Get rid of all-caps speaker headings\n",
    "    return list(zip(*[doc[i:] for i in range(n)]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{(My, name, is),\n",
       " (name, is, Marcus),\n",
       " (is, Marcus, am),\n",
       " (Marcus, am, years),\n",
       " (am, years, old),\n",
       " (years, old, In),\n",
       " (old, In, late),\n",
       " (In, late, t),\n",
       " (late, t, remember),\n",
       " (t, remember, exactly),\n",
       " (remember, exactly, when),\n",
       " (exactly, when, there),\n",
       " (when, there, was),\n",
       " (there, was, a),\n",
       " (was, a, lot),\n",
       " (a, lot, of),\n",
       " (lot, of, fighting),\n",
       " (of, fighting, in),\n",
       " (fighting, in, Goma),\n",
       " (in, Goma, where),\n",
       " (Goma, where, was),\n",
       " (where, was, living),\n",
       " (was, living, with),\n",
       " (living, with, my),\n",
       " (with, my, cousins),\n",
       " (my, cousins, My),\n",
       " (cousins, My, parents),\n",
       " (My, parents, and),\n",
       " (parents, and, sisters),\n",
       " (and, sisters, were),\n",
       " (sisters, were, already),\n",
       " (were, already, killed),\n",
       " (already, killed, by),\n",
       " (killed, by, the),\n",
       " (by, the, Rebels),\n",
       " (the, Rebels, They),\n",
       " (Rebels, They, had),\n",
       " (They, had, attacked),\n",
       " (had, attacked, our),\n",
       " (attacked, our, village),\n",
       " (our, village, one),\n",
       " (village, one, night),\n",
       " (one, night, with),\n",
       " (night, with, guns),\n",
       " (with, guns, and),\n",
       " (guns, and, grenades),\n",
       " (and, grenades, went),\n",
       " (grenades, went, to),\n",
       " (went, to, a),\n",
       " (to, a, rally),\n",
       " (a, rally, in),\n",
       " (rally, in, Goma),\n",
       " (in, Goma, where),\n",
       " (Goma, where, several),\n",
       " (where, several, Government),\n",
       " (several, Government, commanders),\n",
       " (Government, commanders, spoke),\n",
       " (commanders, spoke, and),\n",
       " (spoke, and, said),\n",
       " (and, said, that),\n",
       " (said, that, the),\n",
       " (that, the, President),\n",
       " (the, President, and),\n",
       " (President, and, the),\n",
       " (and, the, were),\n",
       " (the, were, defending),\n",
       " (were, defending, Congo),\n",
       " (defending, Congo, remember),\n",
       " (Congo, remember, one),\n",
       " (remember, one, of),\n",
       " (one, of, the),\n",
       " (of, the, commanders),\n",
       " (the, commanders, who),\n",
       " (commanders, who, spoke),\n",
       " (who, spoke, was),\n",
       " (spoke, was, Chief),\n",
       " (was, Chief, because),\n",
       " (Chief, because, everyone),\n",
       " (because, everyone, cheered),\n",
       " (everyone, cheered, when),\n",
       " (cheered, when, he),\n",
       " (when, he, got),\n",
       " (he, got, up),\n",
       " (got, up, to),\n",
       " (up, to, speak),\n",
       " (to, speak, There),\n",
       " (speak, There, were),\n",
       " (There, were, many),\n",
       " (were, many, soldiers),\n",
       " (many, soldiers, at),\n",
       " (soldiers, at, the),\n",
       " (at, the, rally),\n",
       " (the, rally, about),\n",
       " (rally, about, the),\n",
       " (about, the, same),\n",
       " (the, same, as),\n",
       " (same, as, the),\n",
       " (as, the, number),\n",
       " (the, number, of),\n",
       " (number, of, children),\n",
       " (of, children, in),\n",
       " (children, in, my),\n",
       " (in, my, school),\n",
       " (my, school, Each),\n",
       " (school, Each, family),\n",
       " (Each, family, had),\n",
       " (family, had, to),\n",
       " (had, to, contribute),\n",
       " (to, contribute, a),\n",
       " (contribute, a, cow),\n",
       " (a, cow, or),\n",
       " (cow, or, a),\n",
       " (or, a, goat),\n",
       " (a, goat, or),\n",
       " (goat, or, a),\n",
       " (or, a, child),\n",
       " (a, child, was),\n",
       " (child, was, around),\n",
       " (was, around, ten),\n",
       " (around, ten, years),\n",
       " (ten, years, old),\n",
       " (years, old, then),\n",
       " (old, then, My),\n",
       " (then, My, aunt),\n",
       " (My, aunt, said),\n",
       " (aunt, said, should),\n",
       " (said, should, go),\n",
       " (should, go, since),\n",
       " (go, since, my),\n",
       " (since, my, family),\n",
       " (my, family, was),\n",
       " (family, was, dead),\n",
       " (was, dead, because),\n",
       " (dead, because, of),\n",
       " (because, of, the),\n",
       " (of, the, Rebels),\n",
       " (the, Rebels, followed),\n",
       " (Rebels, followed, the),\n",
       " (followed, the, soldiers)}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grailGrams = set(ngrams(pdf_text, 3))\n",
    "grailGrams"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "for gram in grailGrams:\n",
    "    if sum([word.prob for word in gram]) < -40:\n",
    "        print(gram)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Word Embeddings (Word Vectors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "pdfNouns = [word for word in pdf_text if word.pos_.startswith('N')][:150]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "pdfNounVecs = [word.vector for word in pdfNouns]\n",
    "pdfNounLabels = [word.string.strip() for word in pdfNouns]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(300,)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pdfNounVecs[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "lsa = TruncatedSVD(n_components=2)\n",
    "lsaOut = lsa.fit_transform(pdfNounVecs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xs, ys = lsaOut[:,0], lsaOut[:,1]\n",
    "plt.figure(figsize=[10,10])\n",
    "for i in range(len(xs)): \n",
    "    plt.scatter(xs[i], ys[i])\n",
    "    plt.annotate(pdfNounLabels[i], (xs[i], ys[i]))"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}