Mass-decrypt PGP messages in Thunderbird folders for CLI-based email searchability

pgpgrep.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-

import sys
import subprocess
import argparse
import re
import mailbox
import email.utils
import os
import os.path
import configparser
from os.path import expanduser
from pick import pick

homedir = expanduser("~")

def find_mailboxes():
  config = configparser.ConfigParser()
  config.read(homedir + '/.thunderbird/profiles.ini')
  tbird_profile = config['Profile0']['Path']
  tbird_imap_path = homedir + '/.thunderbird/' + tbird_profile + '/ImapMail/'
  imap_servers = next(os.walk(tbird_imap_path))[1]
  mailboxes = {}
  for x in imap_servers:
    mailboxes[x] = []
    imap_folder = tbird_imap_path + x
    for f in os.listdir(imap_folder):
      file_path = tbird_imap_path + x + '/' + f
      if not os.path.isfile(file_path):
        continue
      if not os.path.splitext(file_path)[1]:
        mailboxes[x].append(file_path)
  return mailboxes

def select_folder(mailboxes):
  prompt = 'Please select an email account: '
  accounts = list(mailboxes.keys())
  account, index = pick(accounts, prompt)
  prompt = 'Please select which folder to search: '
  folders = mailboxes[account]
  folder, index = pick(folders, prompt)
  mbox = mailbox.mbox(folder)
  return folder

def parse_cmd_args():
  description = """
Allows one to extract, decrypt and search through OpenPGP messages scattered among regular unencrypted e-mails.
Decrypted messages are simply printed to standard output so you can use GNU coreutils, etc. to locate information.

Examples:
     # Prompt for IMAP folder selection and decrypt all armored PGP message blocks in a user's local Thunderbird profile
     python3 pgpgrep.py thunderbird
"""

  parser = argparse.ArgumentParser(description=description,
                                   formatter_class=argparse.RawTextHelpFormatter)

  subparsers = parser.add_subparsers(help='Subcommands', dest='subcommand')

  thunderbird_parser = subparsers.add_parser('thunderbird', help='Extract OpenPGP armored message blocks from a Thunderbird profile')

  return parser.parse_args()

def extract(data):

  pgp_msgs = re.findall('-----BEGIN PGP MESSAGE-----.*?-----END PGP MESSAGE-----', data, flags=re.DOTALL)
  pgp_signed_msgs = re.findall('-----BEGIN PGP MESSAGE-----.*?-----BEGIN PGP SIGNATURE-----.*?-----END PGP SIGNATURE-----', data, flags=re.DOTALL)
  return list(filter(None,pgp_msgs + pgp_signed_msgs))

def decrypt(ciphertext):

  gpg = subprocess.Popen(['gpg2', '--decrypt', '--quiet'], stdout=subprocess.PIPE, stdin=subprocess.PIPE, stderr=subprocess.PIPE)
  gpg.stdin.write(ciphertext.encode('utf-8'))
  plaintext = gpg.communicate()[0]
  gpg.stdin.close()
  return plaintext.decode('utf-8')

def main():

  args = parse_cmd_args()
  if args.subcommand == 'thunderbird':
    mailboxes = find_mailboxes()
    pgp_msgs = []
    folder = select_folder(mailboxes)
    if not (os.access(folder, os.R_OK)):
      sys.stderr.write('Please confirm that ' + os.path.abspath(folder) + ' exists and is readable.\n')
      sys.exit(1)
  
    mbox = mailbox.mbox(folder)
    print('Reading messages from: ' + folder)

    for message in mbox:
      try:
         body = message.as_string()
      except:
        continue
      pgp_content = extract(body)
      if pgp_content:
        pgp_msgs.append('\n'.join(pgp_content))
      print('PGP messages found: ' + str(len(pgp_msgs)))
      for pgp_msg in pgp_msgs:
        plaintext = decrypt(pgp_msg)
        print(plaintext)
      sys.exit(0)

if __name__ == '__main__':
    main()