~5-a-day/five-a-day-stats/trunk

#!/usr/bin/python

#

# Copyright (C) 2008-2009  Canonical Ltd.

#

# arguments are the mbox file for mailing list file with team members

#

from mailbox import PortableUnixMailbox

from email import message_from_file

from email.errors import MessageParseError

from email.utils import parseaddr

from email.utils import parsedate

import time

import sys

import os

import calendar

MESSAGE_NOT_PARSEABLE = object()

# perhaps you could just get a list of everyone in five-a-day and then find

# their preferred e-mail address, or fallback to display name if it is private

def message_factory(fp):

    try:

        return message_from_file(fp)

    except MessageParseError:

        # Don't return None since that will stop the mailbox iterator.

        return MESSAGE_NOT_PARSEABLE

def scan_bugs(messages, old_msg_id):

    data = []

    msg_id = ""

    new_msg_id_found = False

    for count, message in enumerate(messages):

        # Skip broken messages.

        if message is MESSAGE_NOT_PARSEABLE:

            continue

        msg_id = message['message-id']

        if msg_id == old_msg_id or not old_msg_id:

            new_msg_id_found = True

        # don't re-do bugs over and over again

        if not new_msg_id_found or msg_id == old_msg_id:

            continue

        # Check to ensure it's from a Launchpad bug.

        reply_to = message['reply-to']

        if reply_to is None:

            continue

        reply_name, reply_address = parseaddr(reply_to)

        reply_local_part, reply_domain = reply_address.split('@')

        bug_id = int(reply_local_part)

        # parsedate() returns a tuple - all times are in UTC

        sent_date = parsedate(message['date'])

        # calendar.timegm is seconds since the epoch in UTC

        action_date = int(calendar.timegm(sent_date))

        dummy, sender = parseaddr(message.get('from'))

        payload = message.get_payload()

        subject = message.get('subject', '')

        # Reporting bugs doesn't count as part of five-a-day

        if (reply_domain != 'bugs.launchpad.net' or not reply_local_part.isdigit()) or  \

           '[NEW]' in subject:

            continue

        data.append((msg_id, action_date, sender, bug_id))

    return (data, msg_id)

if __name__ == '__main__':

    if not len(sys.argv) == 2:

        print >> sys.stderr, "Please specify mailbox file."

        sys.exit(1)

    mailbox_file = os.path.expanduser(sys.argv[1])

    if not os.path.exists(mailbox_file):

        print '''%s mailbox file doesn't exist''' % mailbox_file

        sys.exit(1)

    msgidfile_name = mailbox_file+".msgid"

    if os.path.exists(msgidfile_name) and \

       os.stat(msgidfile_name).st_mtime-os.stat(mailbox_file).st_mtime > 60*60:

        sys.exit(0)

    mailbox = PortableUnixMailbox(open(mailbox_file, 'rb'), message_factory)

    if os.path.exists(msgidfile_name):

        old_msg_id = open(msgidfile_name).read().strip()

    else:

        old_msg_id = ""

    (data, last_msg_id) = scan_bugs(mailbox, old_msg_id)

    data_file = open(mailbox_file+".data", 'a')

    for entry in data:

        data_file.write("\"%s\",\"%s\",\"%s\",\"%s\"\n" % (entry[0], entry[1], entry[2], entry[3]))

    data_file.close()

    if os.path.exists(msgidfile_name):

        os.remove(msgidfile_name)

    msgidfile = open(msgidfile_name, "w")

    msgidfile.write(last_msg_id)

    msgidfile.close()