challenging-america-word-gap-prediction/main.ipynb

{
 "metadata": {
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10-final"
  },
  "orig_nbformat": 2,
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3 (ipykernel)",
   "language": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2,
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "dev_data = list()\n",
    "directory = 'test-A'\n",
    "data_path = directory+'/in.tsv'\n",
    "expected_path = directory+'/expected.tsv'\n",
    "out_path = directory+'/out.tsv'\n",
    "\n",
    "with open(data_path, \"r\") as f:\n",
    "    for line in f.readlines():\n",
    "        dev_data.append(line.split('\\t')[-2])\n",
    "\n",
    "# dev_expected = list()\n",
    "# with open(expected_path, \"r\") as f:\n",
    "#     for line in f.readlines():\n",
    "#         dev_expected.append(line.replace('\\n',''))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def interpolate(bigram, unigramCounts, bigramCounts, listOfProb):\n",
    "    lambdaValue = 0.4\n",
    "    word1 = bigram[0]\n",
    "    word2 = bigram[1]\n",
    "    listOfProb[bigram] = (bigramCounts.get(bigram, 0))/(unigramCounts.get(word1, 0)) + (1-lambdaValue)*(unigramCounts.get(word2, 0))/(unigramCounts.get(word1, 0))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calcProbability(bigram, unigramCounts, bigramCounts, listOfProb):\n",
    "    word1 = bigram[0]\n",
    "    word2 = bigram[1]\n",
    "    listOfProb[bigram] = ((bigramCounts.get(bigram, 0))/len(bigramCounts.items()))/((unigramCounts.get(word1, 0))/len(unigramCounts.items()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.tokenize import word_tokenize \n",
    "import re\n",
    "\n",
    "def createBigram(data, expected):\n",
    "   listOfBigrams = []\n",
    "   bigramCounts = {}\n",
    "   unigramCounts = {}\n",
    "\n",
    "   for i in range(len(data)):\n",
    "       tokenized = word_tokenize(data[i])\n",
    "       word = tokenized[-1]\n",
    "       word = word.lower()\n",
    "       word = re.sub('\\W+','', word)\n",
    "       exp = expected[i].lower()\n",
    "       listOfBigrams.append((word, exp))\n",
    "       if (word, exp) in bigramCounts:\n",
    "        bigramCounts[(word, exp)] += 1\n",
    "       else:\n",
    "        bigramCounts[(word, exp)] = 1\n",
    "       if word in unigramCounts:\n",
    "        unigramCounts[word] += 1\n",
    "       else:\n",
    "        unigramCounts[word] = 1\n",
    "        \n",
    "   return listOfBigrams, unigramCounts, bigramCounts\n",
    "\n",
    "def calcBigramProb(listOfBigrams, unigramCounts, bigramCounts):\n",
    "    listOfProb = {}\n",
    "    for bigram in listOfBigrams:\n",
    "        calcProbability(bigram, unigramCounts, bigramCounts, listOfProb)\n",
    "    return listOfProb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "bigrams, uniCounts, biCounts = createBigram(dev_data, dev_expected)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "probs = calcBigramProb(bigrams, uniCounts, biCounts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_results(probs, in_data):\n",
    "    with open(out_path, 'w') as f:\n",
    "        for i in range(len(in_data)):\n",
    "            tokenized = word_tokenize(in_data[i])\n",
    "            word = tokenized[-1]\n",
    "            word = word.lower()\n",
    "            word = re.sub('\\W+','', word)\n",
    "            word_probs = dict(filter(lambda elem: elem[0][0] == word, probs.items()))\n",
    "            word_probs = dict(sorted(word_probs.items(), key=lambda item: item[1], reverse=True))\n",
    "            rest = 1.0 - sum(word_probs.values())\n",
    "            word_probs = list(map(lambda elem: elem[0][1] + \":\"  + '{:.7f}'.format(elem[1]), list(word_probs.items())))\n",
    "            word_probs.append(':'+'{:.7f}'.format(rest))\n",
    "            word_probs = ' '.join(word_probs)\n",
    "            word_probs += '\\n'\n",
    "            f.write(word_probs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "save_results(probs, dev_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ]
}