clippings-syllables-analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "import pylab as pl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from io import open\n",
    "import seaborn as sns\n",
    "import Levenshtein\n",
    "from collections import Counter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "orecs=[]\n",
    "f = open('./data/clippings_analysis.txt')\n",
    "for l in f:\n",
    "    tokens=l.strip().split(' ')\n",
    "    orecs.append((tokens[0], u' '.join(tokens[1:]).strip()))\n",
    "df=pd.DataFrame().from_records(orecs, columns=['clipping','full'])\n",
    "df.to_csv('./data/clippings_analysis.csv', sep='\\t', encoding='utf-8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df_clipping_only=pd.read_csv('./data/only_clippings_gold.phonemes', sep='\\t', skiprows=3, header=None, names=['word','phonemes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_num_syllables(v):\n",
    "    vowels=set(['AA','AE','AH','AO','AW','AY','EH','ER','EY','IH','IY','OW','OY','UH','UW'])\n",
    "    phonemes=v.split(' ')\n",
    "    cv=0\n",
    "    for p in phonemes:\n",
    "        if p in vowels:\n",
    "            cv=cv+1\n",
    "    return cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def clip_helper(v, num_syllables=2, ctype='BACK'):\n",
    "    vowels=set(['AA','AE','AH','AO','AW','AY','EH','ER','EY','IH','IY','OW','OY','UH','UW'])\n",
    "    phonemes=v.split(' ')        \n",
    "    cv=0\n",
    "    output=[]\n",
    "    if ctype=='BACK':\n",
    "        for p in phonemes:\n",
    "            if p in vowels:\n",
    "                cv=cv+1\n",
    "                if cv > num_syllables:\n",
    "                    break\n",
    "            output.append(p)\n",
    "    elif ctype == 'FORE':\n",
    "        phonemes_v=' '.join(phonemes[::-1])\n",
    "        cp = clip_helper(phonemes_v, num_syllables, ctype='BACK')\n",
    "        output = cp.split(' ')[::-1]\n",
    "    return ' '.join(output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def clip(v, num_syllables=2):\n",
    "    vowels=set(['AA','AE','AH','AO','AW','AY','EH','ER','EY','IH','IY','OW','OY','UH','UW'])\n",
    "    phonemes=v.split(' ')\n",
    "    ctype=global_pmf.sample()\n",
    "    if '@' in v: # compound clippins\n",
    "        part1=v.split('@')[0]\n",
    "        part2=v.split('@')[1]\n",
    "        part1_phonemes = part1.strip().split(' ')\n",
    "        part2_phonemes = part2.strip().split(' ')\n",
    "        part1_clip=clip_helper(' '.join(part1_phonemes), num_syllables=1)\n",
    "        part2_clip=clip_helper(' '.join(part2_phonemes), num_syllables=1)\n",
    "        print v, part1_clip, part1_phonemes, part2_clip, part2_phonemes\n",
    "        assert '@' not in part1_clip\n",
    "        assert '@' not in part2_clip\n",
    "\n",
    "        return part1_clip + ' ' + part2_clip\n",
    "    else:\n",
    "        return clip_helper(v, num_syllables, ctype=ctype)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def random_clip(v, num_syllables=2):\n",
    "    vowels=set(['AA','AE','AH','AO','AW','AY','EH','ER','EY','IH','IY','OW','OY','UH','UW'])\n",
    "    v=v.replace('@',' ')\n",
    "    phonemes=v.split(' ')\n",
    "    keep_length=np.random.randint(1, len(v)+1)\n",
    "    return ' '.join(phonemes[:keep_length])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df_clipping_only[\"num\"]=[get_num_syllables(ps) for ps in df_clipping_only.phonemes.values] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df_components_only=pd.read_csv('./data/only_clippings_full_components_gold.phonemes', sep='\\t', skiprows=3, header=None, names=['word','phonemes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df_components_only[\"num\"]=[get_num_syllables(ps) for ps in df_components_only.phonemes.values] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    576.000000\n",
       "mean       3.302083\n",
       "std        1.417553\n",
       "min        1.000000\n",
       "25%        2.000000\n",
       "50%        3.000000\n",
       "75%        4.000000\n",
       "max       14.000000\n",
       "Name: num, dtype: float64"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_components_only.num.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    576.000000\n",
       "mean       1.557292\n",
       "std        0.601606\n",
       "min        1.000000\n",
       "25%        1.000000\n",
       "50%        2.000000\n",
       "75%        2.000000\n",
       "max        4.000000\n",
       "Name: num, dtype: float64"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_clipping_only.num.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class Pmf(Counter):\n",
    "    \"\"\"A Counter with probabilities.\"\"\"\n",
    "\n",
    "    def normalize(self):\n",
    "        \"\"\"Normalizes the PMF so the probabilities add to 1.\"\"\"\n",
    "        total = float(sum(self.values()))\n",
    "        for key in self:\n",
    "            self[key] /= total\n",
    "\n",
    "    def __add__(self, other):\n",
    "        \"\"\"Adds two distributions.\n",
    "\n",
    "        The result is the distribution of sums of values from the\n",
    "        two distributions.\n",
    "\n",
    "        other: Pmf\n",
    "\n",
    "        returns: new Pmf\n",
    "        \"\"\"\n",
    "        pmf = Pmf()\n",
    "        for key1, prob1 in self.items():\n",
    "            for key2, prob2 in other.items():\n",
    "                pmf[key1 + key2] += prob1 * prob2\n",
    "        return pmf\n",
    "\n",
    "    def __hash__(self):\n",
    "        \"\"\"Returns an integer hash value.\"\"\"\n",
    "        return id(self)\n",
    "    \n",
    "    def __eq__(self, other):\n",
    "        return self is other\n",
    "\n",
    "    def render(self):\n",
    "        \"\"\"Returns values and their probabilities, suitable for plotting.\"\"\"\n",
    "        return zip(*sorted(self.items()))\n",
    "    \n",
    "    def sample(self):\n",
    "        keys, vals= zip(*self.items())\n",
    "        return np.random.choice(keys, p=vals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from collections import Counter\n",
    "syllable_counter_clippings=Counter(df_clipping_only.num.values)\n",
    "syllable_counter_clippings_pmf = Pmf(syllable_counter_clippings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "syllable_counter_clippings_pmf.normalize()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_counter_err(c, xlabel, ylabel, topk=10, ymax=0.6, color_dict=None, use_constant_color=False):\n",
    "    import statsmodels.api as sm\n",
    "    z=c.most_common()\n",
    "    print \"Sum\", sum(zip(*z)[1])\n",
    "    sfixes=list(zip(*z)[0])\n",
    "    svals=list(zip(*z)[1])\n",
    "    evals = sm.stats.multinomial_proportions_confint(np.array(svals))\n",
    "    print \"Evals\", evals.shape\n",
    "    nc=Pmf(c)\n",
    "    nc.normalize()\n",
    "    nz=nc.most_common()\n",
    "    nsvals = list(zip(*nz)[1])\n",
    "    lerrs=np.array(nsvals)-evals[:,0] #CI is not necessarily symmetrical\n",
    "    herrs=evals[:,1]-np.array(nsvals)\n",
    "    errs=np.array([lerrs,herrs])\n",
    "    cmap=sns.light_palette(\"purple\", reverse=True, n_colors=25)\n",
    "    if use_constant_color == False:\n",
    "        g=sns.barplot(sfixes[:topk], list(nsvals[:topk]),  capsize=0.2,palette=cmap,**{'yerr':errs[:,:topk]})\n",
    "    else:\n",
    "        g=sns.barplot(sfixes[:topk], list(nsvals[:topk]), palette=color_dict,capsize=0.2,**{'yerr':errs[:,:topk]})\n",
    "    g.set_xticklabels(g.get_xticklabels(), rotation=0, fontsize=14)\n",
    "    g.set_xlabel(xlabel, fontsize=14)\n",
    "    g.set_ylabel(ylabel, fontsize=14)\n",
    "    g.set_ylim((0,ymax))\n",
    "    return g"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sum 576\n",
      "Evals (4, 2)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAESCAYAAAAIfCk9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEmdDAlJUidDQpLUyZCQJHX6/13MVP2vUUuuAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fab61041510>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_counter_err(syllable_counter_clippings, '#(Syllables)', 'Probability')\n",
    "pl.savefig('./figs/clippings_num_syllables.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}