Facebook_message_reverse.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "worthy-screw",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pprint import pprint\n",
    "import xml.etree.ElementTree as ET\n",
    "\n",
    "# Load in html file\n",
    "# tree = ET.parse(\"C:/Users/Malcolm's laptop/Downloads/message_1.html\")\n",
    "tree = ET.parse(\"C:/Users/Malcolm's laptop/Downloads/message_2.html\")\n",
    "root = tree.getroot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "silver-daniel",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3962 items in _4t5n\n",
      "\n",
      "first 5 elements:\n",
      "[{'class': '_7s7q'},\n",
      " {'class': '_7s7q'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'}]\n",
      "\n",
      "last 5 elements:\n",
      "[{'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': '_7s7q'},\n",
      " {'class': '_7s7q'}]\n"
     ]
    }
   ],
   "source": [
    "# Get the <div class=\"_4t5n\"> element\n",
    "_4t5n = [item for item in root.iter() if item.get('class') == '_4t5n'][0]\n",
    "\n",
    "print(f'{len(_4t5n)} items in _4t5n')\n",
    "print(\"\\nfirst 5 elements:\")\n",
    "pprint([child.attrib for child in _4t5n[:5]])\n",
    "print(\"\\nlast 5 elements:\")\n",
    "pprint([child.attrib for child in _4t5n[-5:]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "id": "scheduled-boutique",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "first 5 elements:\n",
      "[{'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'}]\n",
      "\n",
      "last 5 elements:\n",
      "[{'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'}]\n"
     ]
    }
   ],
   "source": [
    "# Get the messages and reverse them\n",
    "messages = list(filter(lambda child: child.get('class') != '_7s7q', _4t5n))[::-1]\n",
    "\n",
    "print(\"\\nfirst 5 elements:\")\n",
    "pprint([child.attrib for child in messages[:5]])\n",
    "print(\"\\nlast 5 elements:\")\n",
    "pprint([child.attrib for child in messages[-5:]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "nutritional-richmond",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add the bottom page number links to the bottom of the messages\n",
    "messages.append(_4t5n[-1])\n",
    "messages.append(_4t5n[-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "id": "white-globe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2 items in _4t5n\n",
      "\n",
      "All elements in _4t5n:\n",
      "[{'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': 'pam _3-95 _2pi0 _2lej uiBoxWhite noborder'},\n",
      " {'class': '_7s7q'},\n",
      " {'class': '_7s7q'}]\n"
     ]
    }
   ],
   "source": [
    "# Remove all messages (<div class=\"_7s7q\"> elements) from _4t5n\n",
    "for child in list(filter(lambda child: child.get('class') != '_7s7q', _4t5n)):\n",
    "    _4t5n.remove(child)\n",
    "    \n",
    "# Remove the bottom page number links from _4t5n\n",
    "_4t5n.remove(_4t5n[-1])\n",
    "_4t5n.remove(_4t5n[-1])\n",
    "print(f'{len(_4t5n)} items in _4t5n')\n",
    "print(\"\\nAll elements in _4t5n:\")\n",
    "pprint([child.attrib for child in messages[-5:]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "id": "minute-hanging",
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid syntax (<ipython-input-173-cf1e32d18c25>, line 9)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-173-cf1e32d18c25>\"\u001b[1;36m, line \u001b[1;32m9\u001b[0m\n\u001b[1;33m    pprint([child.attrib for child in _4t5n[-5:]])1\u001b[0m\n\u001b[1;37m                                                  ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid syntax\n"
     ]
    }
   ],
   "source": [
    "# Add reversed messages and bottom page links back to _4t5n\n",
    "for child in messages:\n",
    "    _4t5n.append(child)\n",
    "\n",
    "print(f'{len(_4t5n)} items in _4t5n')\n",
    "print(\"\\nfirst 5 elements:\")\n",
    "pprint([child.attrib for child in _4t5n[:5]])\n",
    "print(\"\\nlast 5 elements:\")\n",
    "pprint([child.attrib for child in _4t5n[-5:]])1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "considerable-surprise",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "messages/inbox/saadhassan_12hym6rvtg/message_1_old_at_top.html\n",
      "messages/inbox/saadhassan_12hym6rvtg/message_2_old_at_top.html\n",
      "messages/inbox/saadhassan_12hym6rvtg/message_1_old_at_top.html\n",
      "messages/inbox/saadhassan_12hym6rvtg/message_2_old_at_top.html\n"
     ]
    }
   ],
   "source": [
    "# Update links to point to the new filename\n",
    "import re\n",
    "for item in _4t5n.iter():\n",
    "    if '_42ft' in item.attrib.get('class', []):\n",
    "        item.attrib['href'] = re.sub(r'message_(\\d+).html', r'message_\\1_old_at_top.html', item.attrib['href'])\n",
    "        print(item.attrib['href'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "pleasant-holder",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Output to file\n",
    "# tree.write(\"C:/Users/Malcolm's laptop/Downloads/message_1_old_at_top.html\", method=\"html\")\n",
    "tree.write(\"C:/Users/Malcolm's laptop/Downloads/message_2_old_at_top.html\", method=\"html\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}