Build OpenSubtitles Corpus.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## OpenSubtitles corpus \n",
    "The following code was used to extract Polish OpenSubtitles corpus.\n",
    "It consists of ~775 milion tokens and ~143 milion sentences, vastly dialogues which makes it best suited for human to machine interface applications.\n",
    "\n",
    "It is build using a data from Open Parallel Corpis: http://opus.lingfil.uu.se/OpenSubtitles2016.php \n",
    "\n",
    "**IMPORTANT**: If you use the OpenSubtitle corpus: \n",
    "Please, add a link to http://www.opensubtitles.org/ to your website and to your reports and publications produced with the data! \n",
    "\n",
    "You need at least 30GB of free storage to process the data.\n",
    "\n",
    "The firs step it to prepare a data folder."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Truncated data/OpenSubtitles2016.txt\n"
     ]
    }
   ],
   "source": [
    "%%bash\n",
    "mkdir -p data\n",
    "truncatefile > data/OpenSubtitles2016.txt 2&> /dev/null\n",
    "echo \"Truncated data/OpenSubtitles2016.txt\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next step is to download the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "import codecs\n",
    "import lxml.etree as etree\n",
    "import os\n",
    "import regex\n",
    "from tqdm import tqdm\n",
    "from urllib.request import urlretrieve\n",
    "import subprocess\n",
    "\n",
    "class DLProgress(tqdm):\n",
    "    \"\"\" A class to display progressbar...\"\"\"\n",
    "    last_block = 0\n",
    "\n",
    "    def hook(self, block_num=1, block_size=1, total_size=None):\n",
    "        self.total = total_size\n",
    "        self.update((block_num - self.last_block) * block_size)\n",
    "        self.last_block = block_num\n",
    "\n",
    "def download_dump(arch_uri=\"http://opus.lingfil.uu.se/download.php?f=OpenSubtitles2016/\", \n",
    "                  file=\"pl.tar.gz\"):\n",
    "    \"\"\"Download archive file\"\"\"\n",
    "    datafile=os.path.join(\"data\", file)\n",
    "    if not (os.path.isfile(datafile) or os.path.isfile(datafile[:-4])):\n",
    "        with DLProgress(unit='B', unit_scale=True, miniters=1, desc=file) as pbar:\n",
    "            urlretrieve(arch_uri+file, datafile, pbar.hook)\n",
    "    print (\"Downloading DONE\")\n",
    " \n",
    "    return datafile\n",
    "\n",
    "def extract(filepath):\n",
    "    \"\"\"extract using subprocess\"\"\"\n",
    "    bashCommand = [\"tar\",'-xvf', filepath, '-C','data']\n",
    "    try:\n",
    "        output = subprocess.check_output(bashCommand, stderr=subprocess.STDOUT)\n",
    "    except subprocess.CalledProcessError as pserror: \n",
    "        print (pserror.output)\n",
    "    else:\n",
    "        print (\"DONE {}\".format(output))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading DONE\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "IOPub data rate exceeded.\n",
      "The notebook server will temporarily stop sending output\n",
      "to the client in order to avoid crashing it.\n",
      "To change this limit, set the config variable\n",
      "`--NotebookApp.iopub_data_rate_limit`.\n"
     ]
    }
   ],
   "source": [
    "subtitles=download_dump(arch_uri=\"http://opus.lingfil.uu.se/download.php?f=OpenSubtitles2016/\", file=\"pl.tar.gz\")\n",
    "extract(subtitles)"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cleaning up the data\n",
    "\n",
    "When I initially parsed the data and build a word2vec out of it I realized that each and every common word has a number of transcription variants in the vocabulary. Some were typos but most were due to improper character encoding. The letter ą was encoded as š or ŕ, the letter ć as č, æ or ă, letter ę as ê or ć etc..\n",
    "\n",
    "I analyzed it further and found that about 1000 files contained improperly encoded characters. I plotted the histogram of number of improperly encoded tokens in the file.\n",
    "\n",
    "![image.png](attachment:image.png)\n",
    "\n",
    "Files with las than 20 improperly encoded tokens are probably containing a foreign name from some other language like \"Siniškievicia\", \"Siniša\", \"Pašić\", \"Miško\", \"Blatište\", \"Tidža\" etc. or a single typo with no influence on the whole corpus. \n",
    "\n",
    "Files with 20 to 30 improperly encoded tokens are probably having mixed encoding, like having dialogues in multiple languages or composed of two translations each encoded in different encoding.\n",
    "\n",
    "Files with more than 50 to 80 tokens containign foreing characters are improperly encoded in total. \n",
    "\n",
    "A solution to that is to completely discard the files with more than 20 of improperly encoded tokens. It's 278 files, less than 0.07% of all files and over 56000 improperly encoded tokens."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Removing improperly encoded files\n"
     ]
    }
   ],
   "source": [
    "%%bash\n",
    "echo \"Removing improperly encoded files\"\n",
    "while read i; do\n",
    "  rm -f $i\n",
    "done < OpenSubtitle-discard-files.txt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Buildign corpus file\n",
    "\n",
    "The following code builds txt corpus file. In the target file each line is a new sentence with tokens separated by spaces. The basic punctuation is retained. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 0 ns, sys: 0 ns, total: 0 ns\n",
      "Wall time: 18.4 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "from io import StringIO, BytesIO\n",
    "def build_OpenSubtitlesCorpus(filepath, txt_file=\"data/OpenSubtitles2016.txt\"):\n",
    "    with codecs.open(txt_file, 'a', 'utf-8') as fout:\n",
    "        \"\"\"\n",
    "        Extract txt corpus from xml files\n",
    "        \"\"\"\n",
    "        # parse XML\n",
    "        tree= etree.parse(filepath)\n",
    "        s= ((etree.tostring(tree.getroot()) ))\n",
    "        s=BytesIO(s)\n",
    "\n",
    "        doc=[]\n",
    "        sent=[]\n",
    "        \"\"\"Iterate over xml word and sentence tags and build text corpus\"\"\"\n",
    "        for action, elem in etree.iterparse(s, tag=['w','s']):\n",
    "            \n",
    "            # append all words in sentence\n",
    "            if elem.text is not None and elem.text!=\"\":\n",
    "                sent.append(elem.text)\n",
    "                \n",
    "            # when the sentence is finished\n",
    "            if (action=='end' and elem.tag=='s'):\n",
    "                if sent:\n",
    "                    doc.append(\" \".join(sent))\n",
    "                sent=[]\n",
    "\n",
    "        doc=\"\".join(doc)\n",
    "        doc = regex.sub(u\"[^ \\n\\p{Latin}\\-'‘’.?!]\", \" \", doc) # clean text\n",
    "        doc = regex.sub(u\" - \", \" \", doc) # Dialog hypens\n",
    "        doc = regex.sub(u\" *\\n *\", \"\\n\", doc) # Squeeze spaces\n",
    "        doc = regex.sub(u\" +\", \" \", doc) # Squeeze spaces        \n",
    "\n",
    "        # wrote to file\n",
    "        fout.write(doc.lower()+\"\\n\")\n",
    "\n",
    "                \n",
    "#build_OpenSubtitlesCorpus(\"data/OpenSubtitles2016/xml/pl/1998/568102/3380400.xml.gz\", txt_file=\"OS-test.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DONE\n",
      "CPU times: user 1h 10min 44s, sys: 16.6 s, total: 1h 11min\n",
      "Wall time: 1h 11min 12s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "for root, subdirs, files in (os.walk(\"data/OpenSubtitles2016/xml/pl\")):\n",
    "    for filename in (files):\n",
    "        file_path = os.path.join(root, filename)\n",
    "        build_OpenSubtitlesCorpus(file_path)\n",
    "        \n",
    "print (\"DONE\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:py3]",
   "language": "python",
   "name": "conda-env-py3-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}