gmail_chat.py

import email
import email.parser
import imaplib
import getpass


def imap_login(username=None, password=None):
    """
    login and return the connection
    """
    username = username or raw_input("Gmail address: ")
    password = password or getpass.getpass()

    sock = imaplib.IMAP4_SSL("imap.gmail.com", 993)
    sock.login(username, password)

    return sock 

def search_chats(sock):
    """
    Return a list of msg_id's in your Chats label
    """
    ok, _ = sock.select("[Gmail]/Chats", readonly=True)
    if ok == 'NO':
        sock.select("[Google Mail]/Chats", readonly=True)

    ok, [ids] = sock.search(None, '(ALL)')
    ids = ids.split()
    return ids


def get_message(msg_id):
    """
    returns the content of an individual chat
    """

    resp, data = sock.fetch(msg_id, '(RFC822)')

    msg = email.parser.Parser().parsestr(data[0][1])
    for part in msg.walk():
        if part.get_content_type() != 'text/html':
            continue
    return part.get_payload(decode=True)

if __name__ == '__main__':

    sock = imap_login()

    chat_ids = search_chats(sock)

    for id in chat_ids:
        payload = get_message(id)
        print id, len(payload)
        open('data/%05d.html' % int(id), 'w').write(payload)