pochta_serezhki_inbox.py

# Задача: http://forum.ru-board.com/topic.cgi?forum=8&bm=1&topic=82194
# Почта: username@yandex.ru / password

# TODO: после завершение - Убрать в настройках разрешение доступа по портальному паролю:
# https://mail.yandex.ru/?uid=1130000012843751#setup/client

# Использованы материалы: https://eax.me/python-imap/

# Внешние зависимости:
# pip install pandas
# pip install tqdm


# TODO: Добавить удаление дубликатов

import imaplib
import email
from email.message import EmailMessage
import pandas as pd
from tqdm import tqdm

print("\nStarted\n")

imap = imaplib.IMAP4_SSL('imap.yandex.ru')
imap.login('username@yandex.ru', 'password')

# print(imap.list())
imap.select("INBOX")

result, mail_ids = imap.search(None, 'ALL')
mail_ids = mail_ids[0].split()
# print(mail_ids)

contacts = pd.DataFrame(columns=['name', 'email', 'is_parsed'])

for mail_ids in tqdm(mail_ids):
    status, data = imap.fetch(mail_ids, '(RFC822)')
    msg = email.message_from_bytes(data[0][1], _class=EmailMessage)
    # print(msg.items())
    is_parsed = False
    try:
        name, addr = email.utils.parseaddr(msg['From'])
        is_parsed = True
    except Exception as e:
        print(f"\nCannot parse msg['From']: '{msg['From']}'\n")
        name = msg['From']
        addr = ''
    if is_parsed:
        addr = addr.strip().lower()
        if name:
            name = name.strip()
            decoded_string, charset = email.header.decode_header(name)[0]
            if charset is not None:
                try:
                    name = decoded_string.decode(charset)
                except UnicodeDecodeError:
                    print(f"\nCannot decode name: '{name}'\n")
                    name = ''
            else:
                name = decoded_string
    # print(f"\n{name}\t{addr}\t{is_parsed}")
    contacts = contacts.append({'name': name, 'email': addr, 'is_parsed': is_parsed}, ignore_index=True)

with pd.ExcelWriter('pochta_serezhki_inbox.xlsx') as writer:
    contacts.to_excel(writer, sheet_name='contacts', startrow=0, startcol=0)

print("\nCompleted\n")