{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# hindinballs:\n",
    "### Contributed by Praveen Yadav, [GitHub Repo](https://github.com/fnc11/nball4treehindi)\n",
    "## Resources Used:\n",
    "Files used for Hindi data generation are taken from this [github repo](https://bitbucket.org/sivareddyg/python-hindi-wordnet/src/master) which mainly took data from [IIT Bombay University](http://www.cfilt.iitb.ac.in/).  \n",
    "You need to download w2v from this [website](https://fasttext.cc/docs/en/crawl-vectors.html).\n",
    "\n",
    "\n",
    "# Experiment 1:  Training and evaluating nball embeddings\n",
    "## Experiment 1.1: Training nball embeddings\n",
    "* Please also go through this [Informative Report](https://drive.google.com/file/d/1ZZXAsNJxBQygkfmtakVvHw2gkLCan0rR/view?usp=sharing) on how Hindi Data is structure and how to process it to use it for this experiment or you can also go through data_preprocessing and data_generation notebooks for Hindi language."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "% you need to create an empty file nball.txt for output\n",
    "\n",
    "$ python nball.py --train_nball /Users/<user-name>/data/nball.txt --w2v /Users/<user-name>/data/cc.hi.300.vec  --ws_child /Users/<user-name>/data/wordSenseChildren.txt  --ws_catcode /Users/<user-name>/data/glove/catCodes.txt  --log log.txt\n",
    "% --train_nball: output file of nball embeddings\n",
    "% --w2v: file of pre-trained word embeddings\n",
    "% --ws_child: file of parent-children relations among word-senses\n",
    "% --ws_catcode: file of the parent location code of a word-sense in the tree structure\n",
    "% --log: log file, shall be located in the same directory as the file of nball embeddings\n",
    "```"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![exp1.1.jpeg](img/exp1.1.jpeg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The training process can take around 3 days. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Experiment 1.2: Checking whether tree structures are perfectly embedded into word-embeddings\n",
    "* main input is the output directory of nballs created in Experiment 1.1\n",
    "* shell command for running the nball construction and training process"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "$ python nball.py --zero_energy <output-path> --ball <output-file> --ws_child /Users/<user-name>/data/wordSenseChildren.txt\n",
    "% --zero_energy <output-path> : output path of the nballs of Experiment 1.1, e.g. ```/Users/<user-name>/data/data_out```\n",
    "% --ball <output-file> : the name of the output nball-embedding file\n",
    "% --ws_child /Users/<user-name>/data/wordSenseChildren.txt: file of parent-children relations among word-senses\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The checking process can take a very long time around 3-4 hours.\n",
    "* result\n",
    "\n",
    "If zero-energy is achieved, a big nball-embedding file will be created ```<output-path>/<output-file>```\n",
    "otherwise, failed relations and word-senses will be printed.\n",
    "\n",
    "** Test result at Ubuntu platform:\n",
    "![exp1.2.jpeg](img/exp1.2.jpeg)\n",
    " \n",
    "- [nball embeddings with 67152 balls](https://drive.google.com/open?id=1d-D7AF9rl2g_QFAGLD-m3N0DT_5-uZLS)\n",
    "- [nball.txt file](https://drive.google.com/open?id=1JWNuc2eBTWDrbG1MCdHlWtxenGVKX8to) "
   ]
  },
   {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table style=\"width:100%\">\n",
    "  <tr>\n",
    "      <td colspan=\"1\" style=\"text-align:left;background-color:#0071BD;color:white\">\n",
    "        <a rel=\"license\" href=\"http://creativecommons.org/licenses/by-nc/4.0/\">\n",
    "            <img alt=\"Creative Commons License\" style=\"border-width:0;float:left;padding-right:10pt\"\n",
    "                 src=\"https://i.creativecommons.org/l/by-nc/4.0/88x31.png\" />\n",
    "        </a>\n",
    "        &copy; T. Dong, C. Bauckhage<br/>\n",
    "        Licensed under a \n",
    "        <a rel=\"license\" href=\"http://creativecommons.org/licenses/by-nc/4.0/\" style=\"color:white\">\n",
    "            CC BY-NC 4.0\n",
    "        </a>.\n",
    "      </td>\n",
    "      <td colspan=\"2\" style=\"text-align:left;background-color:#66A5D1\">\n",
    "          <b>Acknowledgments:</b>\n",
    "          This material was prepared within the project\n",
    "          <a href=\"http://www.b-it-center.de/b-it-programmes/teaching-material/p3ml/\" style=\"color:black\">\n",
    "              P3ML\n",
    "          </a> \n",
    "          which is funded by the Ministry of Education and Research of Germany (BMBF)\n",
    "          under grant number 01/S17064. The authors gratefully acknowledge this support.\n",
    "      </td>\n",
    "  </tr>\n",
    "</table>"
   ]
  }
 ],
 "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
}