exploring an gmail mbox using python

mbox.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# read mbox file, generate some stats\n",
    "import os\n",
    "import re\n",
    "import mailbox\n",
    "\n",
    "mboxFile = \"my.mbox\"\n",
    "mbox = open(mboxFile, \"r\")\n",
    "\n",
    "# read mbox file\n",
    "def readMbox(mbox):\n",
    "    lines = mbox.readlines()\n",
    "    mbox.close()\n",
    "    return lines\n",
    "\n",
    "# get email addresses\n",
    "def getEmails(lines):\n",
    "    emails = []\n",
    "    for line in lines:\n",
    "        if line.startswith(\"From:\"):\n",
    "            regex = r\"[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\\.[a-zA-Z0-9-.]+\"\n",
    "            match = re.search(regex, line)\n",
    "            if match is not None:\n",
    "                email = match.group()\n",
    "                emails.append(email)\n",
    "    return emails\n",
    "\n",
    "# get email domains\n",
    "def getEmailDomains(emails):\n",
    "    domains = []\n",
    "    for email in emails:\n",
    "        try:\n",
    "            domain = email.strip().split(\"@\")[1].lower()\n",
    "        except IndexError:\n",
    "            print(f\"Invalid email address: {email}\")\n",
    "        domains.append(domain)\n",
    "    return domains\n",
    "\n",
    "\n",
    "emails = getEmails(readMbox(mbox))\n",
    "domains = getEmailDomains(emails)\n",
    "\n",
    "print(f\"Total emails: {len(emails)}\")\n",
    "print(f\"Total domains: {len(domains)}\")\n",
    "print(f\"Unique domains: {len(set(domains))}\")\n",
    "print(f\"Top 10 domains: {sorted(set(domains), key=domains.count, reverse=True)[:10]}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Export emails to text files\n",
    "email_dir = \"./myemails\"\n",
    "if not os.path.exists(email_dir):\n",
    "    os.makedirs(email_dir)\n",
    "\n",
    "def get_body(message):\n",
    "    if message.is_multipart():\n",
    "        for part in message.get_payload():\n",
    "            if part.get_content_type() == 'text/plain':\n",
    "                return part.get_payload()\n",
    "            if part.get_content_type() == 'text/html':\n",
    "                return part.get_payload()\n",
    "            #base64\n",
    "            if part.get_content_type() == 'application/octet-stream':\n",
    "                # decode base64\n",
    "                return part.get_payload()\n",
    "    else:\n",
    "        return message.get_payload()\n",
    "\n",
    "def write_email(message, i):\n",
    "    with open(f\"{email_dir}/email_{i}.txt\", \"w\") as f:\n",
    "        f.write(f\"From: {message['From']}\\n\")\n",
    "        f.write(f\"To: {message['To']}\\n\")\n",
    "        f.write(f\"Subject: {message['Subject']}\\n\")\n",
    "        f.write(f\"Date: {message['Date']}\\n\")\n",
    "        f.write(f\"Body: {get_body(message)}\\n\")\n",
    "\n",
    "mbox = mailbox.mbox(mboxFile)\n",
    "for i, message in enumerate(mbox):\n",
    "    write_email(message, i)"
   ]
  }
 ],
 "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}