#! /home/tgolden/bin/python
import os, sys
from mailbox import PortableUnixMailbox

local_domains = ["timgolden.me.uk", "goodtoread.org", "weekdaymasses.org.uk", "westpark-club.org.uk", "sabren.com"]
mail_root = os.path.join (os.path.expanduser ("~"), "mail")
folders = ['Personal', 'Python', 'Websites']
##mailboxes = ['personal', 'python', 'websites', 'Westpark', 'goodtoread.org', 'weekdaymasses.org.uk', 'work', 'IN-archive', 'Sent', 'winsys']

addresses = set ()
for folder in folders:
  folder_path = os.path.join (mail_root, folder)
  if not os.path.isdir (folder_path): 
    print folder_path, "not found"
    continue
  
  for mailbox in os.listdir (folder_path):
    mailbox_path = os.path.join (folder_path, mailbox)
    print mailbox_path
    
    mbox = PortableUnixMailbox (open (mailbox_path, 'U'))
    for mail in mbox:
      name, address = mail.getaddr ("From")
      if address:
        addresses.add (address)
      name, address = mail.getaddr ("To")
      if address:
        addresses.add (address)

    for domain in local_domains:
      addresses = set (a for a in addresses if domain not in a.lower ())

f = open (os.path.join (os.path.expanduser ("~"), ".procmail_whitelist"), "w")
try:
  f.write ("\n".join (addresses))
finally:
  f.close ()