#!/usr/bin/python

#This code is public domain.
#http://keithdevens.com/weblog/archive/2004/May/03/mail2blog

import email,poplib,time,sys,xmlrpclib,os,re

email_user='a dedicated mailbox, such as foo@example.com'
email_pass='password'
blog_user='username'
blog_pass='password'
xmlrpc_server = 'the location of his MT XML-RPC server'
image_dir = '/path/to/where/images/get/saved/to/'
slugsep = ',' #his posts have slugs - the subject of the e-mail is dual-purposed to contain the title of the post as well as the slug for the URL
from_address = 'his main e-mail address' #make sure that the e-mail comes from this address. This is at least a *tiny* sanity check, though not real security
blogid = 1 #his MT blog id

sys.stderr = sys.stdout #send tracebacks to browser (used when I was testing this as a CGI)

### Functions ###
def writefile(filename, content):
	if not os.path.exists(filename):
		f = file(filename, 'wb')
		f.write(content)
		f.close()
### End Functions ###

mailbox = poplib.POP3('localhost');
mailbox.user(email_user)
mailbox.pass_(email_pass)

num = len(mailbox.list()[1]) #get number of messages
messages = []
for n in range(num):
	#get raw message
	message = email.message_from_string('\n'.join(mailbox.retr(n+1)[1]))
	if message['from'].find(from_address) != -1:
		messages.append(message)
	mailbox.dele(n+1) #delete the message after we're done with it, and hope we don't fail below :)

mailbox.quit()

posts = []
for message in messages:
	foo = message['subject'].split(slugsep, 1)
	if len(foo) == 1: #no separator
		title = foo[0]
		slug  = re.sub(r'(\W+)','-',re.match(r'^(?:\W+)?(.+?)(?:\W+)?$',title).group(1))
	else:
		slug  = foo[0]
		title = foo[1]

	if not title: #posts must have some title
		continue

	#this parses the date from the e-mail, and makes sure the DST setting is derived
	# from the current DST setting (assigning -1 to the DST setting,
	# which is position 8 in the tuple, does that)
	d = list(email.Utils.parsedate(message['date'])); d[8] = -1; d = tuple(d)
	post = {
		'title' : title.strip(),
		'mt_keywords' : slug.strip(),
		'dateCreated' : xmlrpclib.DateTime(time.mktime(d)),
		'images': []
	}
	if message.is_multipart():
		payload = message.get_payload()
		for p in payload:
			if not p.is_multipart(): #there should be no nesting beyond one level
				if p.get_main_type() == 'text': #body of the message
					post['description'] = p.get_payload(decode=True).strip()
				elif p.get_main_type() == 'image':
					post['images'].append({'image':p.get_payload(decode=True),'filename':p.get_filename()})
	else:
		post['description'] = message.get_payload(decode=True).strip()

	posts.append(post)

if posts:
	#save each image to the image_dir
	for post in posts:
		for image in post['images']:
			writefile(image_dir+image['filename'], image['image'])

	#now that I have the posts, post them to movable-type using XML-RPC
	server = xmlrpclib.ServerProxy(xmlrpc_server)
	for post in posts:
		del(post['images'])
		server.metaWeblog.newPost(blogid,blog_user,blog_pass,post,True)