create_trove_dataframe.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create a DataFrame from the Trove dataset\n",
    "\n",
    "### Preprocessing of the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from glob import glob\n",
    "import jellyfish\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "import re"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Obtaining the data\n",
    "\n",
    "You can download the whole Trove dataset using:\n",
    "\n",
    "`wget --recursive --no-parent http://overproof.projectcomputing.com/datasets/`\n",
    "\n",
    "Change the parent directory `overproof.projectcomputing.com` to a friendlier name `trove_overproof`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-01-15 09:35:32--  http://overproof.projectcomputing.com/datasets/\n",
      "Resolving overproof.projectcomputing.com (overproof.projectcomputing.com)... 148.251.129.168\n",
      "Connecting to overproof.projectcomputing.com (overproof.projectcomputing.com)|148.251.129.168|:80... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     873  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (28.7 MB/s) - ‘overproof.projectcomputing.com/datasets/index.html’ saved [873]\n",
      "\n",
      "Loading robots.txt; please ignore errors.\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/robots.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/robots.txt’\n",
      "\n",
      "overproof.projectco     [ <=>                ]  12.58K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (140 MB/s) - ‘overproof.projectcomputing.com/robots.txt’ saved [12881]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/BLDemo/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/BLDemo/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     306  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (24.3 MB/s) - ‘overproof.projectcomputing.com/datasets/BLDemo/index.html’ saved [306]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/bl/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/bl/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     559  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (35.5 MB/s) - ‘overproof.projectcomputing.com/datasets/bl/index.html’ saved [559]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/dataset1/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     323  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (30.8 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/index.html’ saved [323]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/dataset2/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset2/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     350  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (41.7 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset2/index.html’ saved [350]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/dataset3/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset3/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     350  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (41.7 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset3/index.html’ saved [350]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/OVERPROOF_EVALUATION_DATA_README.html\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 153 [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/OVERPROOF_EVALUATION_DATA_README.html’\n",
      "\n",
      "overproof.projectco 100%[===================>]     153  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:35:35 (36.5 MB/s) - ‘overproof.projectcomputing.com/datasets/OVERPROOF_EVALUATION_DATA_README.html’ saved [153/153]\n",
      "\n",
      "--2020-01-15 09:35:35--  http://overproof.projectcomputing.com/datasets/BLDemo/ocrtest-output.zip\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 32270325 (31M) [application/zip]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/BLDemo/ocrtest-output.zip’\n",
      "\n",
      "overproof.projectco 100%[===================>]  30.77M   345KB/s    in 1m 42s  \n",
      "\n",
      "2020-01-15 09:37:17 (309 KB/s) - ‘overproof.projectcomputing.com/datasets/BLDemo/ocrtest-output.zip’ saved [32270325/32270325]\n",
      "\n",
      "--2020-01-15 09:37:17--  http://overproof.projectcomputing.com/datasets/bl/blogEntryForBL.html\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 13095 (13K) [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/bl/blogEntryForBL.html’\n",
      "\n",
      "overproof.projectco 100%[===================>]  12.79K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:17 (227 MB/s) - ‘overproof.projectcomputing.com/datasets/bl/blogEntryForBL.html’ saved [13095/13095]\n",
      "\n",
      "--2020-01-15 09:37:17--  http://overproof.projectcomputing.com/datasets/bl/overProofPerformance1.png\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 105666 (103K) [image/png]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/bl/overProofPerformance1.png’\n",
      "\n",
      "overproof.projectco 100%[===================>] 103.19K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:17 (165 MB/s) - ‘overproof.projectcomputing.com/datasets/bl/overProofPerformance1.png’ saved [105666/105666]\n",
      "\n",
      "--2020-01-15 09:37:17--  http://overproof.projectcomputing.com/datasets/bl/overProofPerformance2.png\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 21136 (21K) [image/png]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/bl/overProofPerformance2.png’\n",
      "\n",
      "overproof.projectco 100%[===================>]  20.64K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:17 (83.3 MB/s) - ‘overproof.projectcomputing.com/datasets/bl/overProofPerformance2.png’ saved [21136/21136]\n",
      "\n",
      "--2020-01-15 09:37:17--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]  12.93K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:18 (180 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/index.html’ saved [13241]\n",
      "\n",
      "--2020-01-15 09:37:18--  http://overproof.projectcomputing.com/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     417  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:18 (18.9 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/index.html’ saved [417]\n",
      "\n",
      "--2020-01-15 09:37:18--  http://overproof.projectcomputing.com/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: unspecified [text/html]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/index.html’\n",
      "\n",
      "overproof.projectco     [ <=>                ]     417  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:18 (36.2 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/index.html’ saved [417]\n",
      "\n",
      "--2020-01-15 09:37:18--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1842.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 244028 (238K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1842.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 238.31K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:18 (274 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1842.txt’ saved [244028/244028]\n",
      "\n",
      "--2020-01-15 09:37:18--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1843.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 371464 (363K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1843.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 362.76K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:37:19 (218 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1843.txt’ saved [371464/371464]\n",
      "\n",
      "--2020-01-15 09:37:19--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1844.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 232697 (227K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1844.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 227.24K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:20 (220 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1844.txt’ saved [232697/232697]\n",
      "\n",
      "--2020-01-15 09:37:20--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1845.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 110181 (108K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1845.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 107.60K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:20 (254 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1845.txt’ saved [110181/110181]\n",
      "\n",
      "--2020-01-15 09:37:20--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1846.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 48299 (47K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1846.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  47.17K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:21 (147 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1846.txt’ saved [48299/48299]\n",
      "\n",
      "--2020-01-15 09:37:21--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1847.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 85051 (83K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1847.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  83.06K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:21 (160 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1847.txt’ saved [85051/85051]\n",
      "\n",
      "--2020-01-15 09:37:21--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1848.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 95206 (93K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1848.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  92.97K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:37:21 (182 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1848.txt’ saved [95206/95206]\n",
      "\n",
      "--2020-01-15 09:37:21--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1849.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 225456 (220K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1849.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 220.17K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:22 (186 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1849.txt’ saved [225456/225456]\n",
      "\n",
      "--2020-01-15 09:37:22--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1850.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 93909 (92K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1850.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  91.71K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:23 (162 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1850.txt’ saved [93909/93909]\n",
      "\n",
      "--2020-01-15 09:37:23--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1852.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 184927 (181K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1852.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 180.59K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:24 (277 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1852.txt’ saved [184927/184927]\n",
      "\n",
      "--2020-01-15 09:37:24--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1853.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 148610 (145K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1853.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 145.13K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:26 (159 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1853.txt’ saved [148610/148610]\n",
      "\n",
      "--2020-01-15 09:37:26--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1854.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 348351 (340K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1854.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 340.19K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:37:27 (218 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1854.txt’ saved [348351/348351]\n",
      "\n",
      "--2020-01-15 09:37:27--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1855.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 426869 (417K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1855.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 416.86K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:37:28 (190 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1855.txt’ saved [426869/426869]\n",
      "\n",
      "--2020-01-15 09:37:28--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1856.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 410081 (400K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1856.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 400.47K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:37:29 (213 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1856.txt’ saved [410081/410081]\n",
      "\n",
      "--2020-01-15 09:37:29--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1857.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 223964 (219K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1857.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 218.71K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:30 (165 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1857.txt’ saved [223964/223964]\n",
      "\n",
      "--2020-01-15 09:37:30--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1858.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 233644 (228K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1858.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 228.17K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:37:30 (198 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1858.txt’ saved [233644/233644]\n",
      "\n",
      "--2020-01-15 09:37:30--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1859.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 505094 (493K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1859.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 493.26K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:32 (179 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1859.txt’ saved [505094/505094]\n",
      "\n",
      "--2020-01-15 09:37:32--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1860.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 376704 (368K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1860.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 367.88K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:33 (143 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1860.txt’ saved [376704/376704]\n",
      "\n",
      "--2020-01-15 09:37:33--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1861.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 703581 (687K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1861.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 687.09K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:35 (256 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1861.txt’ saved [703581/703581]\n",
      "\n",
      "--2020-01-15 09:37:35--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1862.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 479885 (469K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1862.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 468.64K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:36 (181 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1862.txt’ saved [479885/479885]\n",
      "\n",
      "--2020-01-15 09:37:36--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1863.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 687450 (671K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1863.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 671.34K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:37 (227 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1863.txt’ saved [687450/687450]\n",
      "\n",
      "--2020-01-15 09:37:37--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1864.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 517126 (505K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1864.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 505.01K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:39 (151 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1864.txt’ saved [517126/517126]\n",
      "\n",
      "--2020-01-15 09:37:39--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1865.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 544127 (531K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1865.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 531.37K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:40 (179 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1865.txt’ saved [544127/544127]\n",
      "\n",
      "--2020-01-15 09:37:40--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1866.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 625976 (611K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1866.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 611.30K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:41 (218 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1866.txt’ saved [625976/625976]\n",
      "\n",
      "--2020-01-15 09:37:41--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1867.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1175148 (1.1M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1867.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.12M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:37:44 (300 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1867.txt’ saved [1175148/1175148]\n",
      "\n",
      "--2020-01-15 09:37:44--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1868.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 727177 (710K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1868.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 710.13K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:46 (228 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1868.txt’ saved [727177/727177]\n",
      "\n",
      "--2020-01-15 09:37:46--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1869.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 773077 (755K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1869.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 754.96K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:48 (219 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1869.txt’ saved [773077/773077]\n",
      "\n",
      "--2020-01-15 09:37:48--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1870.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 830176 (811K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1870.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 810.72K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:37:50 (225 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1870.txt’ saved [830176/830176]\n",
      "\n",
      "--2020-01-15 09:37:50--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1871.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1437970 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1871.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.37M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:37:54 (280 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1871.txt’ saved [1437970/1437970]\n",
      "\n",
      "--2020-01-15 09:37:54--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1872.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 628566 (614K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1872.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 613.83K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:55 (221 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1872.txt’ saved [628566/628566]\n",
      "\n",
      "--2020-01-15 09:37:55--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1873.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 774984 (757K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1873.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 756.82K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:57 (228 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1873.txt’ saved [774984/774984]\n",
      "\n",
      "--2020-01-15 09:37:57--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1874.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 616514 (602K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1874.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 602.06K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:37:59 (223 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1874.txt’ saved [616514/616514]\n",
      "\n",
      "--2020-01-15 09:37:59--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1875.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 774565 (756K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1875.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 756.41K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:38:00 (219 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1875.txt’ saved [774565/774565]\n",
      "\n",
      "--2020-01-15 09:38:00--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1876.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 735020 (718K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1876.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 717.79K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:38:02 (223 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1876.txt’ saved [735020/735020]\n",
      "\n",
      "--2020-01-15 09:38:02--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1877.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 708645 (692K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1877.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 692.04K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:38:04 (222 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1877.txt’ saved [708645/708645]\n",
      "\n",
      "--2020-01-15 09:38:04--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1878.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 630020 (615K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1878.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 615.25K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:38:05 (243 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1878.txt’ saved [630020/630020]\n",
      "\n",
      "--2020-01-15 09:38:05--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1879.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 582128 (568K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1879.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 568.48K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:38:07 (219 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1879.txt’ saved [582128/582128]\n",
      "\n",
      "--2020-01-15 09:38:07--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1880.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 927234 (906K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1880.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 905.50K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:38:09 (254 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1880.txt’ saved [927234/927234]\n",
      "\n",
      "--2020-01-15 09:38:09--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1881.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1015692 (992K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1881.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 991.89K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:38:11 (260 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1881.txt’ saved [1015692/1015692]\n",
      "\n",
      "--2020-01-15 09:38:11--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1882.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1023337 (999K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1882.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 999.35K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:38:13 (267 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1882.txt’ saved [1023337/1023337]\n",
      "\n",
      "--2020-01-15 09:38:13--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1883.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1009293 (986K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1883.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 985.64K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:38:16 (269 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1883.txt’ saved [1009293/1009293]\n",
      "\n",
      "--2020-01-15 09:38:16--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1884.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1442360 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1884.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.38M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:38:20 (278 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1884.txt’ saved [1442360/1442360]\n",
      "\n",
      "--2020-01-15 09:38:20--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1885.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1331694 (1.3M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1885.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.27M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:38:23 (270 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1885.txt’ saved [1331694/1331694]\n",
      "\n",
      "--2020-01-15 09:38:23--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1886.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2278094 (2.2M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1886.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.17M   291KB/s    in 7.6s    \n",
      "\n",
      "2020-01-15 09:38:31 (292 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1886.txt’ saved [2278094/2278094]\n",
      "\n",
      "--2020-01-15 09:38:31--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1887.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1861240 (1.8M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1887.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.77M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:38:35 (293 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1887.txt’ saved [1861240/1861240]\n",
      "\n",
      "--2020-01-15 09:38:35--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1888.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1871214 (1.8M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1888.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.78M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:38:39 (290 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1888.txt’ saved [1871214/1871214]\n",
      "\n",
      "--2020-01-15 09:38:39--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1889.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1507273 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1889.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.44M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:38:42 (311 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1889.txt’ saved [1507273/1507273]\n",
      "\n",
      "--2020-01-15 09:38:42--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1890.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1590971 (1.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1890.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.52M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:38:47 (334 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1890.txt’ saved [1590971/1590971]\n",
      "\n",
      "--2020-01-15 09:38:47--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1891.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1997423 (1.9M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1891.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.90M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:38:53 (336 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1891.txt’ saved [1997423/1997423]\n",
      "\n",
      "--2020-01-15 09:38:53--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1892.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2251001 (2.1M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1892.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.15M   382KB/s    in 6.0s    \n",
      "\n",
      "2020-01-15 09:38:59 (367 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1892.txt’ saved [2251001/2251001]\n",
      "\n",
      "--2020-01-15 09:38:59--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1893.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1564868 (1.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1893.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.49M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:39:03 (347 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1893.txt’ saved [1564868/1564868]\n",
      "\n",
      "--2020-01-15 09:39:03--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1894.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1341607 (1.3M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1894.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.28M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:39:07 (272 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1894.txt’ saved [1341607/1341607]\n",
      "\n",
      "--2020-01-15 09:39:07--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1895.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1565394 (1.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1895.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.49M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:39:10 (321 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1895.txt’ saved [1565394/1565394]\n",
      "\n",
      "--2020-01-15 09:39:10--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1896.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2257003 (2.2M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1896.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.15M   338KB/s    in 6.6s    \n",
      "\n",
      "2020-01-15 09:39:17 (332 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1896.txt’ saved [2257003/2257003]\n",
      "\n",
      "--2020-01-15 09:39:17--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1897.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2067019 (2.0M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1897.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.97M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:39:24 (307 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1897.txt’ saved [2067019/2067019]\n",
      "\n",
      "--2020-01-15 09:39:24--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1898.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2233123 (2.1M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1898.txt’\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "overproof.projectco 100%[===================>]   2.13M   148KB/s    in 9.5s    \n",
      "\n",
      "2020-01-15 09:39:33 (230 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1898.txt’ saved [2233123/2233123]\n",
      "\n",
      "--2020-01-15 09:39:33--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1899.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2540008 (2.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1899.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.42M   461KB/s    in 5.5s    \n",
      "\n",
      "2020-01-15 09:39:39 (451 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1899.txt’ saved [2540008/2540008]\n",
      "\n",
      "--2020-01-15 09:39:39--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1900.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 3081095 (2.9M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1900.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.94M   427KB/s    in 8.3s    \n",
      "\n",
      "2020-01-15 09:39:47 (362 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1900.txt’ saved [3081095/3081095]\n",
      "\n",
      "--2020-01-15 09:39:47--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1901.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 8897884 (8.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1901.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   8.49M   436KB/s    in 23s     \n",
      "\n",
      "2020-01-15 09:40:10 (384 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1901.txt’ saved [8897884/8897884]\n",
      "\n",
      "--2020-01-15 09:40:10--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1902.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 3073853 (2.9M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1902.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.93M   516KB/s    in 7.0s    \n",
      "\n",
      "2020-01-15 09:40:17 (426 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1902.txt’ saved [3073853/3073853]\n",
      "\n",
      "--2020-01-15 09:40:17--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1903.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2696093 (2.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1903.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.57M   506KB/s    in 6.4s    \n",
      "\n",
      "2020-01-15 09:40:23 (412 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1903.txt’ saved [2696093/2696093]\n",
      "\n",
      "--2020-01-15 09:40:23--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1904.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1268933 (1.2M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1904.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.21M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:40:27 (210 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1904.txt’ saved [1268933/1268933]\n",
      "\n",
      "--2020-01-15 09:40:27--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1905.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1479056 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1905.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.41M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:40:30 (257 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1905.txt’ saved [1479056/1479056]\n",
      "\n",
      "--2020-01-15 09:40:30--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1906.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1512486 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1906.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.44M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:40:34 (285 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1906.txt’ saved [1512486/1512486]\n",
      "\n",
      "--2020-01-15 09:40:34--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1907.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1887428 (1.8M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1907.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.80M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:40:38 (280 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1907.txt’ saved [1887428/1887428]\n",
      "\n",
      "--2020-01-15 09:40:38--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1908.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1551175 (1.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1908.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.48M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:40:44 (296 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1908.txt’ saved [1551175/1551175]\n",
      "\n",
      "--2020-01-15 09:40:44--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1909.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1646839 (1.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1909.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.57M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:40:47 (313 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1909.txt’ saved [1646839/1646839]\n",
      "\n",
      "--2020-01-15 09:40:47--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1910.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2090272 (2.0M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1910.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.99M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:40:53 (330 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1910.txt’ saved [2090272/2090272]\n",
      "\n",
      "--2020-01-15 09:40:53--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1911.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1896963 (1.8M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1911.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.81M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:40:59 (281 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1911.txt’ saved [1896963/1896963]\n",
      "\n",
      "--2020-01-15 09:40:59--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1912.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1406050 (1.3M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1912.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.34M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:41:03 (302 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1912.txt’ saved [1406050/1406050]\n",
      "\n",
      "--2020-01-15 09:41:03--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1913.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1632766 (1.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1913.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.56M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:41:06 (272 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1913.txt’ saved [1632766/1632766]\n",
      "\n",
      "--2020-01-15 09:41:06--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1914.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1332908 (1.3M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1914.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.27M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:41:09 (322 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1914.txt’ saved [1332908/1332908]\n",
      "\n",
      "--2020-01-15 09:41:09--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1915.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2003832 (1.9M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1915.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.91M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:41:14 (301 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1915.txt’ saved [2003832/2003832]\n",
      "\n",
      "--2020-01-15 09:41:14--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1916.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 2109822 (2.0M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1916.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   2.01M   315KB/s    in 6.3s    \n",
      "\n",
      "2020-01-15 09:41:21 (328 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1916.txt’ saved [2109822/2109822]\n",
      "\n",
      "--2020-01-15 09:41:21--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1917.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 913914 (892K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1917.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 892.49K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:41:23 (222 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1917.txt’ saved [913914/913914]\n",
      "\n",
      "--2020-01-15 09:41:23--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1918.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 918211 (897K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1918.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 896.69K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:41:25 (227 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1918.txt’ saved [918211/918211]\n",
      "\n",
      "--2020-01-15 09:41:25--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1919.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 736193 (719K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1919.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 718.94K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:41:27 (213 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1919.txt’ saved [736193/736193]\n",
      "\n",
      "--2020-01-15 09:41:27--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1920.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 563627 (550K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1920.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 550.42K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:41:28 (276 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1920.txt’ saved [563627/563627]\n",
      "\n",
      "--2020-01-15 09:41:28--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1921.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 726760 (710K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1921.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 709.73K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:41:31 (317 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1921.txt’ saved [726760/726760]\n",
      "\n",
      "--2020-01-15 09:41:31--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1922.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1502336 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1922.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.43M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:41:35 (299 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1922.txt’ saved [1502336/1502336]\n",
      "\n",
      "--2020-01-15 09:41:35--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1923.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1110122 (1.1M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1923.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.06M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:41:37 (244 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1923.txt’ saved [1110122/1110122]\n",
      "\n",
      "--2020-01-15 09:41:37--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1924.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1560673 (1.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1924.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.49M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:41:42 (317 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1924.txt’ saved [1560673/1560673]\n",
      "\n",
      "--2020-01-15 09:41:42--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1925.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1575113 (1.5M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1925.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.50M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:41:49 (336 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1925.txt’ saved [1575113/1575113]\n",
      "\n",
      "--2020-01-15 09:41:49--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1926.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1645917 (1.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1926.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.57M  --.-KB/s    in 0.006s  \n",
      "\n",
      "2020-01-15 09:41:52 (259 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1926.txt’ saved [1645917/1645917]\n",
      "\n",
      "--2020-01-15 09:41:52--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1927.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1671984 (1.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1927.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.59M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:41:58 (293 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1927.txt’ saved [1671984/1671984]\n",
      "\n",
      "--2020-01-15 09:41:58--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1928.txt\n",
      "Connecting to overproof.projectcomputing.com (overproof.projectcomputing.com)|148.251.129.168|:80... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 843737 (824K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1928.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 823.96K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:42:00 (227 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1928.txt’ saved [843737/843737]\n",
      "\n",
      "--2020-01-15 09:42:00--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1929.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 664968 (649K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1929.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 649.38K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:02 (313 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1929.txt’ saved [664968/664968]\n",
      "\n",
      "--2020-01-15 09:42:02--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1930.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 687782 (672K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1930.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 671.66K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:03 (272 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1930.txt’ saved [687782/687782]\n",
      "\n",
      "--2020-01-15 09:42:03--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1931.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 406808 (397K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1931.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 397.27K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:04 (208 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1931.txt’ saved [406808/406808]\n",
      "\n",
      "--2020-01-15 09:42:04--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1932.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 575889 (562K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1932.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 562.39K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:42:05 (214 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1932.txt’ saved [575889/575889]\n",
      "\n",
      "--2020-01-15 09:42:05--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1933.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 587816 (574K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1933.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 574.04K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:07 (237 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1933.txt’ saved [587816/587816]\n",
      "\n",
      "--2020-01-15 09:42:07--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1934.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1225856 (1.2M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1934.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.17M  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:42:09 (366 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1934.txt’ saved [1225856/1225856]\n",
      "\n",
      "--2020-01-15 09:42:09--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1935.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1639668 (1.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1935.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.56M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:42:13 (333 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1935.txt’ saved [1639668/1639668]\n",
      "\n",
      "--2020-01-15 09:42:13--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1936.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 630118 (615K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1936.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 615.35K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:15 (317 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1936.txt’ saved [630118/630118]\n",
      "\n",
      "--2020-01-15 09:42:15--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1937.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 594288 (580K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1937.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 580.36K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:16 (358 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1937.txt’ saved [594288/594288]\n",
      "\n",
      "--2020-01-15 09:42:16--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1938.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1664272 (1.6M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1938.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.59M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:42:20 (371 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1938.txt’ saved [1664272/1664272]\n",
      "\n",
      "--2020-01-15 09:42:20--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1939.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 3908641 (3.7M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1939.txt’\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "overproof.projectco 100%[===================>]   3.73M   527KB/s    in 7.9s    \n",
      "\n",
      "2020-01-15 09:42:28 (484 KB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1939.txt’ saved [3908641/3908641]\n",
      "\n",
      "--2020-01-15 09:42:28--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1940.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1419786 (1.4M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1940.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.35M  --.-KB/s    in 0.005s  \n",
      "\n",
      "2020-01-15 09:42:32 (271 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1940.txt’ saved [1419786/1419786]\n",
      "\n",
      "--2020-01-15 09:42:32--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1941.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 740506 (723K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1941.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 723.15K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:42:34 (219 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1941.txt’ saved [740506/740506]\n",
      "\n",
      "--2020-01-15 09:42:34--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1942.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 766232 (748K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1942.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 748.27K  --.-KB/s    in 0.003s  \n",
      "\n",
      "2020-01-15 09:42:36 (232 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1942.txt’ saved [766232/766232]\n",
      "\n",
      "--2020-01-15 09:42:36--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1943.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 179264 (175K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1943.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 175.06K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:42:36 (223 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1943.txt’ saved [179264/179264]\n",
      "\n",
      "--2020-01-15 09:42:36--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1944.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 321354 (314K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1944.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 313.82K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:37 (168 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1944.txt’ saved [321354/321354]\n",
      "\n",
      "--2020-01-15 09:42:37--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1945.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1096307 (1.0M) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1945.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]   1.04M  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:42:41 (258 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1945.txt’ saved [1096307/1096307]\n",
      "\n",
      "--2020-01-15 09:42:41--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1946.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 98497 (96K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1946.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  96.19K  --.-KB/s    in 0.001s  \n",
      "\n",
      "2020-01-15 09:42:41 (154 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1946.txt’ saved [98497/98497]\n",
      "\n",
      "--2020-01-15 09:42:41--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1947.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 72070 (70K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1947.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  70.38K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:41 (159 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1947.txt’ saved [72070/72070]\n",
      "\n",
      "--2020-01-15 09:42:41--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1948.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 59758 (58K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1948.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  58.36K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:42 (155 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1948.txt’ saved [59758/59758]\n",
      "\n",
      "--2020-01-15 09:42:42--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1949.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 39050 (38K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1949.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  38.13K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:42 (157 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1949.txt’ saved [39050/39050]\n",
      "\n",
      "--2020-01-15 09:42:42--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1950.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 41011 (40K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1950.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  40.05K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:42 (206 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1950.txt’ saved [41011/41011]\n",
      "\n",
      "--2020-01-15 09:42:42--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1951.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 28758 (28K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1951.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  28.08K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:43 (145 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1951.txt’ saved [28758/28758]\n",
      "\n",
      "--2020-01-15 09:42:43--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1952.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 74580 (73K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1952.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  72.83K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:43 (148 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1952.txt’ saved [74580/74580]\n",
      "\n",
      "--2020-01-15 09:42:43--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 57956 (57K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  56.60K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:43 (198 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt’ saved [57956/57956]\n",
      "\n",
      "--2020-01-15 09:42:43--  http://overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1954.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 90795 (89K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1954.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>]  88.67K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-01-15 09:42:44 (206 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1954.txt’ saved [90795/90795]\n",
      "\n",
      "--2020-01-15 09:42:44--  http://overproof.projectcomputing.com/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/allArticles.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1018734 (995K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/allArticles.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 994.86K  --.-KB/s    in 0.004s  \n",
      "\n",
      "2020-01-15 09:42:49 (239 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/allArticles.txt’ saved [1018734/1018734]\n",
      "\n",
      "--2020-01-15 09:42:49--  http://overproof.projectcomputing.com/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/allArticles.txt\n",
      "Reusing existing connection to overproof.projectcomputing.com:80.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 365931 (357K) [text/plain]\n",
      "Saving to: ‘overproof.projectcomputing.com/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/allArticles.txt’\n",
      "\n",
      "overproof.projectco 100%[===================>] 357.35K  --.-KB/s    in 0.002s  \n",
      "\n",
      "2020-01-15 09:42:50 (220 MB/s) - ‘overproof.projectcomputing.com/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples/allArticles.txt’ saved [365931/365931]\n",
      "\n",
      "FINISHED --2020-01-15 09:42:50--\n",
      "Total wall clock time: 7m 18s\n",
      "Downloaded: 129 files, 150M in 3m 16s (781 KB/s)\n"
     ]
    }
   ],
   "source": [
    "!wget --recursive --no-parent http://overproof.projectcomputing.com/datasets/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.rename('overproof.projectcomputing.com', 'trove_overproof')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocessing of the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dictionary with the following structure:\n",
    "# arbitrary_dirname: path_to_txt_files\n",
    "# dirname is an arbitrary name which can be used to filter the DB in later steps\n",
    "\n",
    "dict_dirnames = {\n",
    "    \"trove_dataset_1\": './trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs',\n",
    "    \"trove_dataset_2\": './trove_overproof/datasets/dataset2/rawTextAndHumanCorrectionAndOverproofCorrectionTriples',\n",
    "    \"trove_dataset_3\": './trove_overproof/datasets/dataset3/rawTextAndHumanCorrectionAndOverproofCorrectionTriples'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(columns=[\"filePath\", \"articleId\", \"articleType\", \"year\", \n",
    "                           \"ocrText\", \"humanText\", \"corrected\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_header(line):\n",
    "    re_header = r'^\\*\\$\\*OVERPROOF\\*\\$\\*\\s*([0-9]+)\\s+year\\s+([0-9]{4})\\s+type\\s(.+)\\s+title\\s+.+$'\n",
    "    if re.match(re_header, line):\n",
    "        articleId, year, articleType = re.match(re_header, line).groups()\n",
    "        return articleId, year, articleType"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_content(article_content):\n",
    "    ocr_line = []\n",
    "    human_line = []\n",
    "    hcorr_line = []\n",
    "    for line in article_content:\n",
    "        line = line.split(\"||@@||\")\n",
    "        if (not line[0].strip() == '') or (not line[1].strip() == ''):\n",
    "            if len(line) == 3:\n",
    "                ocr_line.append(line[0].strip())\n",
    "                human_line.append(line[1].strip())\n",
    "                hcorr_line.append(line[2].strip())\n",
    "            if len(line) == 2:\n",
    "                ocr_line.append(line[0].strip())\n",
    "                human_line.append(line[1].strip())\n",
    "\n",
    "    return \" \".join(ocr_line), \" \".join(human_line), \" \".join(hcorr_line)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30509\n"
     ]
    }
   ],
   "source": [
    "article_content = []\n",
    "year = \"\"\n",
    "file_path = \"\"\n",
    "articleId = \"\"\n",
    "articleType = \"\"\n",
    "processed_content = \"\"\n",
    "article_counter = 0\n",
    "for ldir in dict_dirnames.keys():\n",
    "    list_files = glob(os.path.join(dict_dirnames[ldir], \"*.txt\"))\n",
    "    for lfile in list_files:\n",
    "        file_path = lfile\n",
    "        fio = open(lfile, \"r\")\n",
    "        flines = fio.readlines()\n",
    "        for iline in range(len(flines)):\n",
    "            if flines[iline].startswith('*$*OVERPROOF*$*'):\n",
    "                articleId, year, articleType = process_header(flines[iline])\n",
    "                article_counter += 1\n",
    "            else:\n",
    "                article_content.append(flines[iline])\n",
    "                if ((len(flines) -1 == iline) or (flines[iline + 1].startswith('*$*OVERPROOF*$*'))):\n",
    "                    processed_content = process_content(article_content)\n",
    "\n",
    "                    # Insert into dataframe:\n",
    "                    df.loc[article_counter] = [file_path, \n",
    "                                               articleId,\n",
    "                                               articleType,\n",
    "                                               year,\n",
    "                                               processed_content[0],\n",
    "                                               processed_content[1],\n",
    "                                               processed_content[2]\n",
    "                                              ] \n",
    "                    \n",
    "                    # Clean variables:\n",
    "                    year = \"\"\n",
    "                    articleId = \"\"\n",
    "                    articleType = \"\"\n",
    "                    processed_content = \"\"\n",
    "                    article_content = []\n",
    "\n",
    "print(article_counter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filePath</th>\n",
       "      <th>articleId</th>\n",
       "      <th>articleType</th>\n",
       "      <th>year</th>\n",
       "      <th>ocrText</th>\n",
       "      <th>humanText</th>\n",
       "      <th>corrected</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18378453</td>\n",
       "      <td>Article ILLUSTRATED</td>\n",
       "      <td>1953</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\"...</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18363627</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18366055</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>FIRST CHURCH I SERVICE 1 Presbyterian I ' Anni...</td>\n",
       "      <td>FIRST CHURCH SERVICE Presbyterian Anniversary ...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18386137</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium District...</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium  Arthur ...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18368961</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr Justic...</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr. Justi...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30505</th>\n",
       "      <td>./trove_overproof/datasets/dataset3/rawTextAnd...</td>\n",
       "      <td>4</td>\n",
       "      <td>Article</td>\n",
       "      <td>1921</td>\n",
       "      <td>AIR FORCES SEND SUB TO BTTOM WITH BOMB. Explos...</td>\n",
       "      <td>AIR FORCES SEND SUB TO BOTTOM WITH BOMB. Explo...</td>\n",
       "      <td>AIR FORCES SEND SUB TO BTTOM WITH BOMB. Explos...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30506</th>\n",
       "      <td>./trove_overproof/datasets/dataset3/rawTextAnd...</td>\n",
       "      <td>5</td>\n",
       "      <td>Article</td>\n",
       "      <td>1922</td>\n",
       "      <td>Wh reWYOU 100 Years Ago? alnd S frd Nise. Nw M...</td>\n",
       "      <td>Who Were YOU 100 Years Ago? Blinds Birds of Ni...</td>\n",
       "      <td>Wh reWYOU 100 Years Ago? and S ford Nine. New ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30507</th>\n",
       "      <td>./trove_overproof/datasets/dataset3/rawTextAnd...</td>\n",
       "      <td>6</td>\n",
       "      <td>Article</td>\n",
       "      <td>1922</td>\n",
       "      <td>TRBTY OR WAR, FEAR or LODGE Senator Warns of P...</td>\n",
       "      <td>TRBTY OR WAR, FEAR or LODGE Senator Warns of P...</td>\n",
       "      <td>TRBTY OR WAR, FEAR or LODGE Senator Warns of P...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30508</th>\n",
       "      <td>./trove_overproof/datasets/dataset3/rawTextAnd...</td>\n",
       "      <td>7</td>\n",
       "      <td>Article</td>\n",
       "      <td>1922</td>\n",
       "      <td>RAIDERS IN FREDER IC K LAND HAUL OF $84,000 Wa...</td>\n",
       "      <td>RAIDERS IN FREDERICK LAND HAUL OF $84,000 Watc...</td>\n",
       "      <td>RAIDERS IN FREDERICK LAND HAUL OF $84,000 Watc...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30509</th>\n",
       "      <td>./trove_overproof/datasets/dataset3/rawTextAnd...</td>\n",
       "      <td>8</td>\n",
       "      <td>Article</td>\n",
       "      <td>1922</td>\n",
       "      <td>GUARDED SINCE EATH THREATS British Ambassador'...</td>\n",
       "      <td>GUARDED SINCE DEATH THREATS British Ambassador...</td>\n",
       "      <td>GUARDED SINCE DEATH THREATS British Ambassador...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>30509 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                filePath articleId  \\\n",
       "1      ./trove_overproof/datasets/dataset1/rawTextAnd...  18378453   \n",
       "2      ./trove_overproof/datasets/dataset1/rawTextAnd...  18363627   \n",
       "3      ./trove_overproof/datasets/dataset1/rawTextAnd...  18366055   \n",
       "4      ./trove_overproof/datasets/dataset1/rawTextAnd...  18386137   \n",
       "5      ./trove_overproof/datasets/dataset1/rawTextAnd...  18368961   \n",
       "...                                                  ...       ...   \n",
       "30505  ./trove_overproof/datasets/dataset3/rawTextAnd...         4   \n",
       "30506  ./trove_overproof/datasets/dataset3/rawTextAnd...         5   \n",
       "30507  ./trove_overproof/datasets/dataset3/rawTextAnd...         6   \n",
       "30508  ./trove_overproof/datasets/dataset3/rawTextAnd...         7   \n",
       "30509  ./trove_overproof/datasets/dataset3/rawTextAnd...         8   \n",
       "\n",
       "                articleType  year  \\\n",
       "1      Article ILLUSTRATED   1953   \n",
       "2                   Article  1953   \n",
       "3                   Article  1953   \n",
       "4                   Article  1953   \n",
       "5                   Article  1953   \n",
       "...                     ...   ...   \n",
       "30505               Article  1921   \n",
       "30506               Article  1922   \n",
       "30507               Article  1922   \n",
       "30508               Article  1922   \n",
       "30509               Article  1922   \n",
       "\n",
       "                                                 ocrText  \\\n",
       "1      FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\"...   \n",
       "2      Natural Childbirth Sir,-We nurses have seen fa...   \n",
       "3      FIRST CHURCH I SERVICE 1 Presbyterian I ' Anni...   \n",
       "4      \"Bob\" Lulham's Fight Against Thallium District...   \n",
       "5      DIVORCE Before The Judge In Divorce, Mr Justic...   \n",
       "...                                                  ...   \n",
       "30505  AIR FORCES SEND SUB TO BTTOM WITH BOMB. Explos...   \n",
       "30506  Wh reWYOU 100 Years Ago? alnd S frd Nise. Nw M...   \n",
       "30507  TRBTY OR WAR, FEAR or LODGE Senator Warns of P...   \n",
       "30508  RAIDERS IN FREDER IC K LAND HAUL OF $84,000 Wa...   \n",
       "30509  GUARDED SINCE EATH THREATS British Ambassador'...   \n",
       "\n",
       "                                               humanText  \\\n",
       "1      FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...   \n",
       "2      Natural Childbirth Sir,-We nurses have seen fa...   \n",
       "3      FIRST CHURCH SERVICE Presbyterian Anniversary ...   \n",
       "4      \"Bob\" Lulham's Fight Against Thallium  Arthur ...   \n",
       "5      DIVORCE Before The Judge In Divorce, Mr. Justi...   \n",
       "...                                                  ...   \n",
       "30505  AIR FORCES SEND SUB TO BOTTOM WITH BOMB. Explo...   \n",
       "30506  Who Were YOU 100 Years Ago? Blinds Birds of Ni...   \n",
       "30507  TRBTY OR WAR, FEAR or LODGE Senator Warns of P...   \n",
       "30508  RAIDERS IN FREDERICK LAND HAUL OF $84,000 Watc...   \n",
       "30509  GUARDED SINCE DEATH THREATS British Ambassador...   \n",
       "\n",
       "                                               corrected  \n",
       "1                                                         \n",
       "2                                                         \n",
       "3                                                         \n",
       "4                                                         \n",
       "5                                                         \n",
       "...                                                  ...  \n",
       "30505  AIR FORCES SEND SUB TO BTTOM WITH BOMB. Explos...  \n",
       "30506  Wh reWYOU 100 Years Ago? and S ford Nine. New ...  \n",
       "30507  TRBTY OR WAR, FEAR or LODGE Senator Warns of P...  \n",
       "30508  RAIDERS IN FREDERICK LAND HAUL OF $84,000 Watc...  \n",
       "30509  GUARDED SINCE DEATH THREATS British Ambassador...  \n",
       "\n",
       "[30509 rows x 7 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Add string similarity and length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO: Pandarallel will run on 8 workers.\n",
      "INFO: Pandarallel will use standard multiprocessing data tranfer (pipe) to transfer data between the main process and workers.\n"
     ]
    }
   ],
   "source": [
    "from pandarallel import pandarallel\n",
    "# Initialization\n",
    "pandarallel.initialize()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def distance_via_levenshtein(gs_clean, ocr_clean):\n",
    "    gs_clean = gs_clean.lower()\n",
    "    ocr_clean = ocr_clean.lower()\n",
    "    max_sentlength = max(len(gs_clean), len(ocr_clean))\n",
    "    lev_distance = jellyfish.levenshtein_distance(gs_clean, ocr_clean)\n",
    "    dist_similarity = (max_sentlength - lev_distance) / float(max_sentlength)\n",
    "    return dist_similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['str_similarity'] = df.parallel_apply(lambda row: distance_via_levenshtein(row['ocrText'], row['humanText']), axis=1)\n",
    "df['str_length_humanText'] = df.parallel_apply(lambda row: len(row['humanText']), axis=1)\n",
    "df['str_length_ocrText'] = df.parallel_apply(lambda row: len(row['ocrText']), axis=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5077"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Number of documents that have at least one article with string similarity lower than 0.8:\n",
    "len(df[df['str_similarity'] < 0.8]['articleId'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def quality(similarity):\n",
    "    if similarity > 0.9:\n",
    "        return 1 # good\n",
    "    elif similarity > 0.8:\n",
    "        return 2 # soso\n",
    "    elif similarity > 0.7:\n",
    "        return 3 # bad\n",
    "    return 4 # ugly\n",
    "\n",
    "df['quality_band'] = df[\"str_similarity\"].apply(quality)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2    13953\n",
       "1    11461\n",
       "3     3600\n",
       "4     1495\n",
       "Name: quality_band, dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['quality_band'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OCR and GS string similarity distribution\n",
    "\n",
    "Plot the distribution of sentences according to string similarity between OCR and GS text:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x120c33eb8>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df[['str_similarity']].plot(kind='hist',bins=np.arange(0, 1.1, 0.1),rwidth=0.8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get some examples from each bin:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ocrText</th>\n",
       "      <th>humanText</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21092</th>\n",
       "      <td>A HARBOUIt COIXISIOU. ^ .FERRY CAPTAIN PINED. ...</td>\n",
       "      <td>A HARBOUR COLLISION. FERRY CAPTAIN FINED. At t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11189</th>\n",
       "      <td>BRITISH AND FOREIGN SHIPPINC. LONDON\", Tuesday...</td>\n",
       "      <td>BRITISH AND FOREIGN SHIPPING. LONDON, Tuesday,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4287</th>\n",
       "      <td>SHIPPING. . i ARRIA'ALS.-SEPTEMBRE 13. Ly-eo-A...</td>\n",
       "      <td>SHIPPING. ARRIVALS. — SEPTEMBER 13. Ly-ee-Moon...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1099</th>\n",
       "      <td>! SERGEANT GUA TY OF DESERTION Pound guilty bv...</td>\n",
       "      <td>SERGEANT GUILTY OF DESERTION Found guilty by a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12055</th>\n",
       "      <td>ii METROPOl.lTANiDlSTRIOT COURT.-Flinn.-. {lie...</td>\n",
       "      <td>METROPOLITAN DISTRICT COURT.— FRIDAY. (Before ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15920</th>\n",
       "      <td>HOTEL TRANSFERS, The following transfers of ho...</td>\n",
       "      <td>HOTEL TRANSFERS. The following transfers of ho...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 ocrText  \\\n",
       "21092  A HARBOUIt COIXISIOU. ^ .FERRY CAPTAIN PINED. ...   \n",
       "11189  BRITISH AND FOREIGN SHIPPINC. LONDON\", Tuesday...   \n",
       "4287   SHIPPING. . i ARRIA'ALS.-SEPTEMBRE 13. Ly-eo-A...   \n",
       "1099   ! SERGEANT GUA TY OF DESERTION Pound guilty bv...   \n",
       "12055  ii METROPOl.lTANiDlSTRIOT COURT.-Flinn.-. {lie...   \n",
       "15920  HOTEL TRANSFERS, The following transfers of ho...   \n",
       "\n",
       "                                               humanText  \n",
       "21092  A HARBOUR COLLISION. FERRY CAPTAIN FINED. At t...  \n",
       "11189  BRITISH AND FOREIGN SHIPPING. LONDON, Tuesday,...  \n",
       "4287   SHIPPING. ARRIVALS. — SEPTEMBER 13. Ly-ee-Moon...  \n",
       "1099   SERGEANT GUILTY OF DESERTION Found guilty by a...  \n",
       "12055  METROPOLITAN DISTRICT COURT.— FRIDAY. (Before ...  \n",
       "15920  HOTEL TRANSFERS. The following transfers of ho...  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[(df['str_similarity'] > 0.7) & (df['str_similarity'] <= 0.8)].sample(6).loc[:, ['ocrText', 'humanText']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_pickle(\"db_trove.pkl\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Range: (0.20000000000000004, 0.30000000000000004)\n",
      "Number of articles: 1\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1882.txt\n",
      "\n",
      "OCR text:\n",
      "MVRMPMVJt' Ai'milNEKJù BY HIS WISE. OfftOM OV-JA OW,N; ejQltfllitü.'ÖXDr.H'r.) MUDGEE, Satthudat. Y-uteWiW * XWifP WWWW» VW *\"*» .Jj:tI t¡(-*-- \"? foibiiftht ftßi) ttt Ah Yyni^ ^ (ftiniMW ¿twckcepor tit Hargraves, shot hw -hM*b*w\\ cJcatV í*?oy bud a quai-roi on the iW\\io\\fö WUWtf- wh.qft fto woman lofe her husbilnd iHiijt AV'-fcM H\", iji'cr '.¡twfrifx'-^ residence, laroo lullet] ftn-iy*. Vv*s>ví-.Vvy îp.vrwv'i tho Chinamau, it ont to U'6 )W»ttm J». ï4'S>W*Wft> Wiilin, imothc-r quarrel look placo, (lili) it !*i V(i*ii )bu' ^Juäuuiun 'us-<iiuU«l tho mother. \"'' \" W--?'?'? .*-?.<.*.' - *.--..- .. -- -? tit IWliVe 1 IWi.t*? 1>H JltWSf8 M* J-Mo-TO W raon-uiK to Hold ail mimi-, mvm\\m tâ Wäpmw* w llt, t01*0-,aeni <4 .lUu Ituiwuwa p.W!U«. uuû Uiu ir>ii unaarteateu bv.them. .\n",
      "\n",
      "Human-corrected text:\n",
      "MURDER OF A CHINESE BY HIS WIFE [BY TELEGRAPH] (FROM OUR OWN CORRESPONDENT) MUDGEE, SATURDAY. Yesterday a young woman, who was married a fortnight ago to Ah Yon, a Chinese storekeeper at Hargraves, shot her husband dead. They had a quarrel on the previous evening when the woman left her husband and went to hermother's residence, three miles away. Yesterday morning the Chinaman went to the mother's residence when another quarrel took place, and it is said the Chinaman assaulted the mother. The daughter went to her father's room, took a gun, and shot her husband dead. The news only arrived here last even- ing and Mr. Meares, the coroner, Dr. Rowling, and the police left Mudgee for Hargraves his morning to hold an inquest. Information of the occurrence was at once sent to the Hargraves police and the girl was arrested by them.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.30000000000000004, 0.4)\n",
      "Number of articles: 8\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1923.txt\n",
      "\n",
      "OCR text:\n",
      "NORTH ROCKS FAIR. M tin- iiiiiiunl liir tit Nui Hi Kui.« in \"W .>* Hie lot ni Si boni ni Wth Hie i» mis »on -lorry liuisc-i l> llutclic, 1 II I MMIICI, - l'onln It li null, «m1 Ululer S Smith, I I Munich 2 l'oiilci, to 11 J limul- \\ Hull 1 renier-, ist - list liiicK Muck nell 1, s sini.li, > lot luck I irtler, 1, S Smith 2 rruil hrimci * turnciit G Iliuiie« 1, It I Muller, : Penn li lok I Hull 1 Sulkl runiout \\ Hull, 1 Pcnteisiut, 2 l'uni jump ('. Iliuhe ( liiiiupioti rll.li n nom uiilci II IUIIIIN \\ Hull Uiniiipiim rililmn linn I mir llreiinli l'cmi 11 linns uiul un le r ld », | Smith Llminiilon rib lui MIII.II Imn tin m ut l' lliigln» Ollin nuit- lesillcil III. tailem -Um« nillir illirs W spinnt Cul- un UM II er-- W une \\t lus (ni-, niilor 1 lu 11 nine smith nul I h le Snllli lied It is nillir lilli- UiHiil-i 1! l'mler X e it e h mill < 11 lill» I, It Sim lU nul I I I riinks Heil incluís billilli ip \"M iiiuiis I! lintir-iij HUjcle rue t( nrlin-'furil 11 11.111 t'ili Uni\") c. Soulei ! il Souter - It s\"\"|. (\"|0,t lime linn uiiler Hire (Tluls) Il (1 i in eli Tliriiwing jt wicket: It _i_r, bl<___> ground. W. li. Suiith.\n",
      "\n",
      "Human-corrected text:\n",
      "NORTH ROCKS FAIR. At the annual fair at North Rocks, in aid of the local School of Arts, the awards were: —Lorry horses: G. Hughes, 1: R. J. Maher, 2. Ponies, 13 hands and under: S Smith, 1: E. Moulds, 2. Ponies, to 11.2 hands: A. Bull. 1: Pentergast, 2. 11st hack: Black- well, 1: S. Smith, 2. 10st hack: Lawler, 1: S. Smith 2. Fruit-grower's turnout: G. Hughes, 1: R. J Maher, 2: Pony hack: J. Bull, 1 Sulky turnout A. Hull, 1: Pentergast, 2. Pony jump; G. Hughes Champion ribbon, pony under 14 hands: A. Bull Campion ribbon, brood mare: Bromly. Pony, 13 hands and under led in: E. Smith, Champion rib- bon, single lorry turnout: G. Hughes. Other events resulted as follows:- Boys under 10 yrs W. Sprowles. Girls under 10 years: Winne Atkins. Girls under 13 yrs: Daphne Smith and Elvie Smith tied. Boys under 16 yrs (400 yds): R. Porter. Novice handicap, 100 yds: R. Sheilds and F. E. Franks tied. 100 yds: handicap, all comers: R. Ramsay. Bicycle race (Carlingford League Cycle Club): G. Sonter, 1: R. Sonter, 2: R. Sands, fastest time. Boys under 13 yrs (75yds): D. O'Turell. Throwing at wicket: R. Maher. Stepping ground: W. G. Smith.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.4, 0.5)\n",
      "Number of articles: 36\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1941.txt\n",
      "\n",
      "OCR text:\n",
      "DIVORCE COURT Deciect, nUi (bcfoie Mr Justice Bonney) - lincst willum lux \\ biella lill 1 li duli!. Oioikr Koili 1 inim i Idu Aniu-I I lunn I ninth Once (.-iinUl i Innic H-iiuld (is kau Wiiltir chirrnte nnttlim \\ Join Pottom Dence absolute lnn.il Niiitimbc v lu Kiihlecn Ne \\coinbc Deciros nisi (bcfoie Mi Justice Sheet) - Milschr Ilor-ll/ll <I noun « \\ min linn l I lill Mil II II K Ullin ii lll/nuitll MilLdilh i olmtlri C.romi Mrrt-dilli Um lull«. Druin l lick William Denne Ilumine (Ullin inici v Leila Ilnij Km ilil Air mili Elnvart v hc-iUlu III ni I-ini.es «lei il Alllllir Til Hill iiiil v Ollii Beril Milln ntl len lane HinJii v llirnld lledlci lltnili s Irllltiii Roe Pellcurlnu i lihn Pilli mino Arllmr Uealnald Curtols v Phslll« lllola Mull c innis ilcstliullon older«. Ill Ima Mai Uplon l Birraclmtah Upton Sar iii louise loins i Lt«lic Iinol-nd -lc-nuiMnit Ion\"_\n",
      "\n",
      "Human-corrected text:\n",
      "DIVORCE COURT Decrees nisi (before Mr. Justice Bonney) :— Ernest William Long v Stella Long, Frederick George Koch Einam v Ida Amuel Einam, Francis George Garnett v James Harold Cog- gan, Walter Clarence Bottom v Joyce Bottom. Decree absolute: Ernest Newcombe v Ivy Kathleen Newcombe. Decrees nisi (before Mr. Justice Street) :— Maische Horsitzki (known as Martin Hall) v Lilli Horsitzki, Kathleen Elizabeth Meredith v Stanley George Meredith, Constance Devine v Jack William Devine, Lawrence Clifton Davey v Leila Davey, Ronald Alexander Stewart v Heather Thyra Frances Stewart, Arthur Ray Millward v Olive Beryl Millward, Vera Jane Hughes v Harold Hedley Hughes, Lettitia Rose Pellegrino v John Pellegrino, Arthur Reginald Curtois v Phyllis Thora Mary Curtois. Restitution orders: Thelma May Upton v Barraclough Upton, Sarah Louise Jones v Leslie Rowland Templeton Jones.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.5, 0.6)\n",
      "Number of articles: 194\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1944.txt\n",
      "\n",
      "OCR text:\n",
      "STRATHFIELD WOMAN 100 TO-MORROW Mi«; Emilie I-R-,» of Albeit Road Strathfield will celebiate hei 100th biilhüA*, to monow She came lo AUatmha ni the late sulita and as a named nu se lids de\\oi«.d nrailj ill hu lue lo the eau 01 tnt ¿ick nnu 111«, uUtuic of uiili A n nive oi lA.tuii'-liiit Miss Ka*,« oiutd hu Li tain«; Pt.>iod ul Uie Al lied Hu -jil U AliIboittne and lau Let mi. m mun ol tile L./.mu- llouiv. loi h pt-i-- ut ljuiii, ii ii nnu sub i. l|ltUlll\\ HllUUll ol Uic 1J m mi itli in alislii 1 ¿jUu>ol lui Cllili U lu ¡,he whs niition of a mu in« home ii Mutend vlivS is. ve «ds one of the found crs of the Hciuj G\"oirc league in S*-dne\\ _\n",
      "\n",
      "Human-corrected text:\n",
      "STRATHFIELD WOMAN 100 TO-MORROW Miss Emilie Kaye, of Albert Road Strathfield, will celebrate her 100th birthday tomorrow. She came to Australia in the late sixties, and, as a trained nurse, has devoted nearly all her life to the care of the sick and the welfare of girls. A native of Devonshire, Miss Kaye served her training period at the Al- fred Hospital, Melbourne, and later became matron of the Lazarus Home for Lepers at Long Bay, and subse- quently matron of the Parramatta in- dustrial School for Girls. Later she was matron of a nursing home at Maitland. Miss Kaye was one of the found- ers of the Henry George League in Sydney.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.6000000000000001, 0.7000000000000001)\n",
      "Number of articles: 715\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt\n",
      "\n",
      "OCR text:\n",
      "Mint Building i Sur,--Tb* cía .\\-fint Build- ing in Macquarie Streit- cutir1 hy Franaa Greenaway., is <c ceaurffüiTj prccorocnetf char it m a defigfat sa tàe eye of ant- one wfec fore a. seaae of beaut? aa¿ ¿buga. Its*atnraoi?eaess has,, howeser fresa ruined by the area 5x franc, hacina feeai aa^&aTtedi amf used as a parfc sçgarës. \" Enragáte frew mach mers at- tractive thür charming: oíd fanld ais wouf<£ Ecoic ÉÊ the space ai frone was concerted hito lawas anea gardens; wahi fcrígní ffcwen. Why doesn't the Scare Govern- ' ment use the? cid building to house some of the Law Court»? At present the? are scarcerecS aft over the place, fe seenst st tharne that this fine feuñdóií s&cufif he >jsed for the Housing Ccnmta îfcnvw&iciîcoufd!ea3rï2 ¡removed to sonfe other säe fir Waoflteov rneofccv Mifler/j. Point cr Surr? Hat*. Sjcfeey. EJL. i\n",
      "\n",
      "Human-corrected text:\n",
      "Mint Building Sir, - The old Mint Build- ing in Macquarie Street, built by Francis Greenaway, is so beautifully proportioned that it is a delight to the eye of any- one who has a sense of beauty and design. Its attractiveness has however been ruined by the area in front, having been designated and used as a park- ing area. Imagine how much more at- tractive this charming old build- ing would look if the space in front was converted into lawns and gardens with bright flowers. Why doesn't the State Govern- ment use this old building to house some of the Law Courts. At present they are scattered all over the place. It seems a shame that this fine building should be used for the Housing Commis- sion which could easily be moved to some other site in Woolloo- mooloo, Millers Point or Surry Hills. Sydney E.R.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.7000000000000001, 0.8)\n",
      "Number of articles: 3062\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt\n",
      "\n",
      "OCR text:\n",
      "rchitect For lympics aid £31,665 - ?? MELBOURNE, Friday [r Frank Heath, the tarne architect who Ped the Olympic Sta m for Carlton Oval be t the Olympic Games {f switched to the ^Gi has been awarded »i665 compensation to sum was paid out of »ral Olympic Games \"? to which the Federal «rament contributed 50 «nt and the Victorian «rament and Melbourne Council 25 per cent LHwih had claimed TO. but the Victorian «rament refused to meet ¡ The £31.665 was de «Kd bv an independent Nor Mr j B Tait, Q C «'«nins year, Mr Heath .Paid £8 500 for the '«gun at Carlton This £en deducted from the\n",
      "\n",
      "Human-corrected text:\n",
      "Architect For Olympics paid £31,665  MELBOURNE, Friday Mr Frank Heath, the Melbourne architect who designed ed the Olympic Sta- dium for Carlton Oval be- fore the Olympic Games were switched to the MCG has been awarded £31,665 compensation. The sum was paid out of general Olympic Games \"? to which the Federal Government contributed 50 percent and the Victorian Government and Melbourne Council 25 per cent each. Mr Heath had claimed TO. but the Victorian Government refused to meet this. The £31.665 was de- derived by an independent moderator Mr J B Tait, Q C earlier this year. Mr Heath was paid £8 500 for the work begun at Carlton. This has been deducted from the £31,665.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.8, 0.9)\n",
      "Number of articles: 13053\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt\n",
      "\n",
      "OCR text:\n",
      "FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\" Pclcr Antill ploughed deep into Paddy's River in his Chrysler Plymouth Jr LA Jil ? during {|)c elimination section. Antill has been provisionally placed third. BOG' navigator Miss Marie Higgs pushed hard at the back of Bill McLachlan's Ford Customlinc while it ploughed tnrougti wu\" a boggy crcckcrossing on the elimination section. FINISH' ^ H°'den driven by A. L. Kipling, of South Australia, was the first to cross the line in Driver Avenue yesterday. I lill Jil. Kipling lost his first points on the elimination section. DRINK\" Tom Sn,man (,eft) washed Lex Davi I/I\\II1IV. son wasjj tjie ,jus{ from jjjs throat, HOMF\" Hunlber driver, M. S. Law, was HUÍ IL i greeted by his children. AIITllfiDADH\" ßarrV Gurdon signed auto j MUIVUftHril. graphs at (hc fmish yesterday. L\n",
      "\n",
      "Human-corrected text:\n",
      "FROM RIVER CROSSING TO END OF TRIAL SPLASH: Peler Antill ploughed deep into Paddy's River in his Chrysler Plymouth during the elimination section. Antill has been provisionally placed third. BOG: navigator Miss Marie Higgs pushed hard at the back of Bill McLachlan's Ford Customline while it ploughed tnrough a boggy creekcrossing on the elimination section. FINISH: A Holden driven by A. L. Kipling, of South Australia, was the first to cross the line in Driver Avenue yesterday. Kipling lost his first points on the elimination section. DRINK: Tom Sulman (left) watched Lex Davi- son wash the dust from his throat. HOMF: Humber driver, M. S. Law, was greeted by his children. AUTOGRAPH: ßarry Gordon signed auto- graphs at the finish yesterday.\n",
      "\n",
      "==========\n",
      "\n",
      "Range: (0.9, 1.0)\n",
      "Number of articles: 11218\n",
      "Example: ././trove_overproof/datasets/dataset1/rawTextAndHumanCorrectionPairs/smh1953.txt\n",
      "\n",
      "OCR text:\n",
      "Natural Childbirth Sir,-We nurses have seen far too many women coming into labour wards in fear and trembling, bringine their own pain with them. Doctors are too busy to in- struct or teach expectant moth- ers, and somewhere before their term is near a sympathetic and kindly individual explanation should be available. Not only should they have exercises, but the whole process of childbirth should be explained simply. The ignorance among nurses of the Read methods suggests that the lay population must be doubly ignorant. How far this unnecessary suffering affects the birthrate can only be guessed at. LOUISE'ROBINSON. Bondi.\n",
      "\n",
      "Human-corrected text:\n",
      "Natural Childbirth Sir,-We nurses have seen far too many women coming into labour wards in fear and trembling, bringing bringing their own pain with them. Doctors are too busy to instruct or teach expectant mothers, and somewhere before their term is near a sympathetic and kindly individual explanation should be available. Not only should they have excercises, but the whole process of childbirth should be explained simply. The ignorance among nurse of the Read methods suggests that that the lay population must be doubly ignorant. How far this unnecessary suffering affects the birthrate can only be guessed at. LOUISE ROBINSON. Bondi. \n",
      "\n",
      "==========\n",
      "\n"
     ]
    }
   ],
   "source": [
    "trovedf = df\n",
    "\n",
    "dExamples = dict()\n",
    "for i in np.arange(0.1, 1.1, 0.1):\n",
    "    tempdf = trovedf[(trovedf['str_similarity'] > (i - 0.1)) & (trovedf['str_similarity'] <= i) & (abs(trovedf['ocrText'].str.len() - trovedf['humanText'].str.len()) <= 100)]\n",
    "    if tempdf.shape[0] >= 1:\n",
    "        dExamples[((i - 0.1, i))] = (tempdf.shape[0], tempdf.iloc[0].filePath, tempdf.iloc[0].ocrText, tempdf.iloc[0].humanText)\n",
    "\n",
    "for example in dExamples:\n",
    "    print(\"Range:\", example)\n",
    "    print(\"Number of articles:\", dExamples[example][0])\n",
    "    print(\"Example: ./\" + dExamples[example][1])\n",
    "    print(\"\\nOCR text:\")\n",
    "    print(dExamples[example][2])\n",
    "    print(\"\\nHuman-corrected text:\")\n",
    "    print(dExamples[example][3])\n",
    "    print()\n",
    "    print('==========')\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dictionary lookup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import re\n",
    "import syntok.segmenter as segmenter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filePath</th>\n",
       "      <th>articleId</th>\n",
       "      <th>articleType</th>\n",
       "      <th>year</th>\n",
       "      <th>ocrText</th>\n",
       "      <th>humanText</th>\n",
       "      <th>corrected</th>\n",
       "      <th>str_similarity</th>\n",
       "      <th>str_length_humanText</th>\n",
       "      <th>str_length_ocrText</th>\n",
       "      <th>quality_band</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18378453</td>\n",
       "      <td>Article ILLUSTRATED</td>\n",
       "      <td>1953</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\"...</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...</td>\n",
       "      <td></td>\n",
       "      <td>0.847561</td>\n",
       "      <td>746</td>\n",
       "      <td>820</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18363627</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td></td>\n",
       "      <td>0.964119</td>\n",
       "      <td>641</td>\n",
       "      <td>630</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18366055</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>FIRST CHURCH I SERVICE 1 Presbyterian I ' Anni...</td>\n",
       "      <td>FIRST CHURCH SERVICE Presbyterian Anniversary ...</td>\n",
       "      <td></td>\n",
       "      <td>0.738901</td>\n",
       "      <td>946</td>\n",
       "      <td>832</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18386137</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium District...</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium  Arthur ...</td>\n",
       "      <td></td>\n",
       "      <td>0.493898</td>\n",
       "      <td>2950</td>\n",
       "      <td>2740</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18368961</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr Justic...</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr. Justi...</td>\n",
       "      <td></td>\n",
       "      <td>0.894176</td>\n",
       "      <td>1219</td>\n",
       "      <td>1121</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            filePath articleId  \\\n",
       "1  ./trove_overproof/datasets/dataset1/rawTextAnd...  18378453   \n",
       "2  ./trove_overproof/datasets/dataset1/rawTextAnd...  18363627   \n",
       "3  ./trove_overproof/datasets/dataset1/rawTextAnd...  18366055   \n",
       "4  ./trove_overproof/datasets/dataset1/rawTextAnd...  18386137   \n",
       "5  ./trove_overproof/datasets/dataset1/rawTextAnd...  18368961   \n",
       "\n",
       "            articleType  year  \\\n",
       "1  Article ILLUSTRATED   1953   \n",
       "2               Article  1953   \n",
       "3               Article  1953   \n",
       "4               Article  1953   \n",
       "5               Article  1953   \n",
       "\n",
       "                                             ocrText  \\\n",
       "1  FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\"...   \n",
       "2  Natural Childbirth Sir,-We nurses have seen fa...   \n",
       "3  FIRST CHURCH I SERVICE 1 Presbyterian I ' Anni...   \n",
       "4  \"Bob\" Lulham's Fight Against Thallium District...   \n",
       "5  DIVORCE Before The Judge In Divorce, Mr Justic...   \n",
       "\n",
       "                                           humanText corrected  \\\n",
       "1  FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...             \n",
       "2  Natural Childbirth Sir,-We nurses have seen fa...             \n",
       "3  FIRST CHURCH SERVICE Presbyterian Anniversary ...             \n",
       "4  \"Bob\" Lulham's Fight Against Thallium  Arthur ...             \n",
       "5  DIVORCE Before The Judge In Divorce, Mr. Justi...             \n",
       "\n",
       "   str_similarity  str_length_humanText  str_length_ocrText  quality_band  \n",
       "1        0.847561                   746                 820             2  \n",
       "2        0.964119                   641                 630             1  \n",
       "3        0.738901                   946                 832             3  \n",
       "4        0.493898                  2950                2740             4  \n",
       "5        0.894176                  1219                1121             2  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_pickle(\"db_trove.pkl\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "corrected_cond = (df[\"corrected\"] == '')# & (df[\"str_length_humanText\"] > 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    30301\n",
       "1      208\n",
       "Name: use_corrected, dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"use_corrected\"] = 1\n",
    "df.loc[corrected_cond, 'corrected'] = df.loc[corrected_cond, 'humanText']\n",
    "df.loc[corrected_cond, 'use_corrected'] = 0\n",
    "df['use_corrected'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "ocr_corr_diff = abs(df['ocrText'].str.len() - df['corrected'].str.len())\n",
    "ocr_corr_diff /= np.maximum(df['ocrText'].str.len().values, df['corrected'].str.len().values)\n",
    "ocr_corr_diff *= 100."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20, 10))\n",
    "plt.plot(ocr_corr_diff, c='k')\n",
    "plt.grid()\n",
    "plt.xticks(size=24)\n",
    "plt.yticks(size=24)\n",
    "plt.xlabel(\"ArticleID\", size=32)\n",
    "plt.ylabel(\"|#ocrText - #corrected| (%)\", size=32)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#Articles with character difference lower than 10.00%: 29697, percentage: 97.34%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20, 10))\n",
    "\n",
    "perc_diff = 10.\n",
    "\n",
    "plt.plot(np.sort(ocr_corr_diff), c='k', lw=3)\n",
    "\n",
    "plt.grid()\n",
    "plt.xticks(size=24)\n",
    "plt.yticks(size=24)\n",
    "plt.xlabel(\"#Articles (sorted)\", size=32)\n",
    "plt.ylabel(\"|#ocrText - #corrected| (%)\", size=32)\n",
    "plt.axhline(perc_diff, 0, 1, color='r', ls='--', lw=3)\n",
    "print(\"#Articles with character difference lower than %.2f%%: %i, percentage: %.2f%%\" % (perc_diff, len(ocr_corr_diff[ocr_corr_diff <= perc_diff]), len(ocr_corr_diff[ocr_corr_diff <= perc_diff])/len(ocr_corr_diff)*100.))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# spacy to do most of the pre-processing\n",
    "import spacy\n",
    "# see: https://spacy.io/universe/project/spacy-langdetect\n",
    "from spacy_langdetect import LanguageDetector\n",
    "\n",
    "# preprocessing\n",
    "# Load a spacy model\n",
    "nlp = spacy.load('en_core_web_lg')\n",
    "# spacy_dict will be later used for the \"Dictionary lookup\" evaluation\n",
    "spacy_dict = list(nlp.vocab.strings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dictionary_lookup(myrow, colname=\"corrected\"):\n",
    "    #print(myrow.name, end=\" \")\n",
    "    \n",
    "    sent_list = []\n",
    "    found_dict = []\n",
    "    \n",
    "    for paragraph in segmenter.analyze(myrow[colname]):\n",
    "        for sentence in paragraph:\n",
    "            all_tokens = []\n",
    "            all_txt_dict = []\n",
    "            for token in sentence:\n",
    "                # exactly reproduce the input\n",
    "                # and do not remove \"imperfections\"\n",
    "                # print(token.spacing, token.value, sep='', end='')\n",
    "                all_tokens.append(token.value)\n",
    "                if token.value.lower() in spacy_dict:\n",
    "                    all_txt_dict.append(str(len(token.value)))\n",
    "                else:\n",
    "                    all_txt_dict.append(str(-len(token.value)))\n",
    "            sent_list.append(all_tokens)\n",
    "            found_dict.append(all_txt_dict)\n",
    "            #found_dict.append([])\n",
    "    return sent_list, found_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_pickle(\"db_trove_before_sentencizer.pkl\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO: Pandarallel will run on 8 workers.\n",
      "INFO: Pandarallel will use standard multiprocessing data tranfer (pipe) to transfer data between the main process and workers.\n"
     ]
    }
   ],
   "source": [
    "from pandarallel import pandarallel\n",
    "# Initialization\n",
    "pandarallel.initialize()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['corrected_sentencizer'] = ''\n",
    "df['corrected_dict_lookup'] = ''\n",
    "\n",
    "df['ocr_sentencizer'] = ''\n",
    "df['ocr_dict_lookup'] = ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['corrected_sentencizer'], df['corrected_dict_lookup'] = zip(*df.parallel_apply(dictionary_lookup, args=[\"corrected\"], axis=1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['ocr_sentencizer'], df['ocr_dict_lookup'] = zip(*df.parallel_apply(dictionary_lookup, args=[\"ocrText\"], axis=1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_pickle(\"db_trove_sentence_with_lookup.pkl\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filePath</th>\n",
       "      <th>articleId</th>\n",
       "      <th>articleType</th>\n",
       "      <th>year</th>\n",
       "      <th>ocrText</th>\n",
       "      <th>humanText</th>\n",
       "      <th>corrected</th>\n",
       "      <th>str_similarity</th>\n",
       "      <th>str_length_humanText</th>\n",
       "      <th>str_length_ocrText</th>\n",
       "      <th>quality_band</th>\n",
       "      <th>use_corrected</th>\n",
       "      <th>corrected_sentencizer</th>\n",
       "      <th>corrected_dict_lookup</th>\n",
       "      <th>ocr_sentencizer</th>\n",
       "      <th>ocr_dict_lookup</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18378453</td>\n",
       "      <td>Article ILLUSTRATED</td>\n",
       "      <td>1953</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\"...</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...</td>\n",
       "      <td>FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...</td>\n",
       "      <td>0.847561</td>\n",
       "      <td>746</td>\n",
       "      <td>820</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>[[FROM, RIVER, CROSSING, TO, END, OF, TRIAL, S...</td>\n",
       "      <td>[[4, 5, 8, 2, 3, 2, 5, 6, 1, -5, -6, 8, 4, 4, ...</td>\n",
       "      <td>[[FROM, RIVER, CROSSING, TO, END, OF, TRIÄÜ, I...</td>\n",
       "      <td>[[4, 5, 8, 2, 3, 2, -5, 1, 1, 2, -3, 1, -5, -6...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18363627</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td>Natural Childbirth Sir,-We nurses have seen fa...</td>\n",
       "      <td>0.964119</td>\n",
       "      <td>641</td>\n",
       "      <td>630</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>[[Natural, Childbirth, Sir,-We, nurses, have, ...</td>\n",
       "      <td>[[7, 10, -7, 6, 4, 4, 3, 3, 4, 5, 6, 4, 6, 5, ...</td>\n",
       "      <td>[[Natural, Childbirth, Sir,-We, nurses, have, ...</td>\n",
       "      <td>[[7, 10, -7, 6, 4, 4, 3, 3, 4, 5, 6, 4, 6, 5, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18366055</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>FIRST CHURCH I SERVICE 1 Presbyterian I ' Anni...</td>\n",
       "      <td>FIRST CHURCH SERVICE Presbyterian Anniversary ...</td>\n",
       "      <td>FIRST CHURCH SERVICE Presbyterian Anniversary ...</td>\n",
       "      <td>0.738901</td>\n",
       "      <td>946</td>\n",
       "      <td>832</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>[[FIRST, CHURCH, SERVICE, Presbyterian, Annive...</td>\n",
       "      <td>[[5, 6, 7, 12, 11, 3, 5, 11, 2, 3, 5, 12, 6, 7...</td>\n",
       "      <td>[[FIRST, CHURCH, I, SERVICE, 1, Presbyterian, ...</td>\n",
       "      <td>[[5, 6, 1, 7, 1, 12, 1, 1, 11, 1, 3, 5, -12, 3...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18386137</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium District...</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium  Arthur ...</td>\n",
       "      <td>\"Bob\" Lulham's Fight Against Thallium  Arthur ...</td>\n",
       "      <td>0.493898</td>\n",
       "      <td>2950</td>\n",
       "      <td>2740</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>[[\", Bob, \", Lulham, 's, Fight, Against, Thall...</td>\n",
       "      <td>[[1, 3, 1, -6, 2, 5, 7, 8, 6, 6, 1, 1, 3, 1, 1...</td>\n",
       "      <td>[[\", Bob, \", Lulham, 's, Fight, Against, Thall...</td>\n",
       "      <td>[[1, 3, 1, -6, 2, 5, 7, 8, 8, 8, 1], [4, 5, 6,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>./trove_overproof/datasets/dataset1/rawTextAnd...</td>\n",
       "      <td>18368961</td>\n",
       "      <td>Article</td>\n",
       "      <td>1953</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr Justic...</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr. Justi...</td>\n",
       "      <td>DIVORCE Before The Judge In Divorce, Mr. Justi...</td>\n",
       "      <td>0.894176</td>\n",
       "      <td>1219</td>\n",
       "      <td>1121</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>[[DIVORCE, Before, The, Judge, In, Divorce, ,,...</td>\n",
       "      <td>[[7, 6, 3, 5, 2, 7, 1, 2, 1, 7, -5, 7, 4, 1, 1...</td>\n",
       "      <td>[[DIVORCE, Before, The, Judge, In, Divorce, ,,...</td>\n",
       "      <td>[[7, 6, 3, 5, 2, 7, 1, 2, 7, -5, 7, 4, 1, 1, -...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            filePath articleId  \\\n",
       "1  ./trove_overproof/datasets/dataset1/rawTextAnd...  18378453   \n",
       "2  ./trove_overproof/datasets/dataset1/rawTextAnd...  18363627   \n",
       "3  ./trove_overproof/datasets/dataset1/rawTextAnd...  18366055   \n",
       "4  ./trove_overproof/datasets/dataset1/rawTextAnd...  18386137   \n",
       "5  ./trove_overproof/datasets/dataset1/rawTextAnd...  18368961   \n",
       "\n",
       "            articleType  year  \\\n",
       "1  Article ILLUSTRATED   1953   \n",
       "2               Article  1953   \n",
       "3               Article  1953   \n",
       "4               Article  1953   \n",
       "5               Article  1953   \n",
       "\n",
       "                                             ocrText  \\\n",
       "1  FROM RIVER CROSSING TO END OF TRIÄÜ I ^PI A^H\"...   \n",
       "2  Natural Childbirth Sir,-We nurses have seen fa...   \n",
       "3  FIRST CHURCH I SERVICE 1 Presbyterian I ' Anni...   \n",
       "4  \"Bob\" Lulham's Fight Against Thallium District...   \n",
       "5  DIVORCE Before The Judge In Divorce, Mr Justic...   \n",
       "\n",
       "                                           humanText  \\\n",
       "1  FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...   \n",
       "2  Natural Childbirth Sir,-We nurses have seen fa...   \n",
       "3  FIRST CHURCH SERVICE Presbyterian Anniversary ...   \n",
       "4  \"Bob\" Lulham's Fight Against Thallium  Arthur ...   \n",
       "5  DIVORCE Before The Judge In Divorce, Mr. Justi...   \n",
       "\n",
       "                                           corrected  str_similarity  \\\n",
       "1  FROM RIVER CROSSING TO END OF TRIAL SPLASH: Pe...        0.847561   \n",
       "2  Natural Childbirth Sir,-We nurses have seen fa...        0.964119   \n",
       "3  FIRST CHURCH SERVICE Presbyterian Anniversary ...        0.738901   \n",
       "4  \"Bob\" Lulham's Fight Against Thallium  Arthur ...        0.493898   \n",
       "5  DIVORCE Before The Judge In Divorce, Mr. Justi...        0.894176   \n",
       "\n",
       "   str_length_humanText  str_length_ocrText  quality_band  use_corrected  \\\n",
       "1                   746                 820             2              0   \n",
       "2                   641                 630             1              0   \n",
       "3                   946                 832             3              0   \n",
       "4                  2950                2740             4              0   \n",
       "5                  1219                1121             2              0   \n",
       "\n",
       "                               corrected_sentencizer  \\\n",
       "1  [[FROM, RIVER, CROSSING, TO, END, OF, TRIAL, S...   \n",
       "2  [[Natural, Childbirth, Sir,-We, nurses, have, ...   \n",
       "3  [[FIRST, CHURCH, SERVICE, Presbyterian, Annive...   \n",
       "4  [[\", Bob, \", Lulham, 's, Fight, Against, Thall...   \n",
       "5  [[DIVORCE, Before, The, Judge, In, Divorce, ,,...   \n",
       "\n",
       "                               corrected_dict_lookup  \\\n",
       "1  [[4, 5, 8, 2, 3, 2, 5, 6, 1, -5, -6, 8, 4, 4, ...   \n",
       "2  [[7, 10, -7, 6, 4, 4, 3, 3, 4, 5, 6, 4, 6, 5, ...   \n",
       "3  [[5, 6, 7, 12, 11, 3, 5, 11, 2, 3, 5, 12, 6, 7...   \n",
       "4  [[1, 3, 1, -6, 2, 5, 7, 8, 6, 6, 1, 1, 3, 1, 1...   \n",
       "5  [[7, 6, 3, 5, 2, 7, 1, 2, 1, 7, -5, 7, 4, 1, 1...   \n",
       "\n",
       "                                     ocr_sentencizer  \\\n",
       "1  [[FROM, RIVER, CROSSING, TO, END, OF, TRIÄÜ, I...   \n",
       "2  [[Natural, Childbirth, Sir,-We, nurses, have, ...   \n",
       "3  [[FIRST, CHURCH, I, SERVICE, 1, Presbyterian, ...   \n",
       "4  [[\", Bob, \", Lulham, 's, Fight, Against, Thall...   \n",
       "5  [[DIVORCE, Before, The, Judge, In, Divorce, ,,...   \n",
       "\n",
       "                                     ocr_dict_lookup  \n",
       "1  [[4, 5, 8, 2, 3, 2, -5, 1, 1, 2, -3, 1, -5, -6...  \n",
       "2  [[7, 10, -7, 6, 4, 4, 3, 3, 4, 5, 6, 4, 6, 5, ...  \n",
       "3  [[5, 6, 1, 7, 1, 12, 1, 1, 11, 1, 3, 5, -12, 3...  \n",
       "4  [[1, 3, 1, -6, 2, 5, 7, 8, 8, 8, 1], [4, 5, 6,...  \n",
       "5  [[7, 6, 3, 5, 2, 7, 1, 2, 7, -5, 7, 4, 1, 1, -...  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:ocr-evaluation]",
   "language": "python",
   "name": "conda-env-ocr-evaluation-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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}