This post originated from an RSS feed registered with Python Buzz by Aaron Brady. Original Post: More Handling Mail in Python Feed Title: insom.me.uk Feed URL: http://feeds2.feedburner.com/insommeuk Feed Description: Posts related to using Python. Some tricks and tips, observations, hacks, and the Brand New Things. Latest Python Buzz Posts Latest Python Buzz Posts by Aaron Brady Latest Posts From insom.me.uk

I have updated some hacks filed under mailtools. This includes one new hack, a POP3 mail checker that uses Python for filtering mail.

I used to struggle with the procmail syntax for filtering mail, and now that I'm using PINE on win32 for my primary mail platform, I don't have a good way of checking and filtering mail.

These two combined lead to getmail.py, a Pythonic POP3 Processor (phew!). This simple script stands on the shoulders of the standard library giants, and checks POP3 accounts, downloading them into mbox files.

Aside from being written in Python, it also uses Python as its filtering language. Each mail will call a function of your choice, pass it the account details and the body of the message, and write it out according to the result.

I'd advise looking at the admittedly messy source, but here's a quick example script to use to check a few accounts with lists:

import getmail
import re
import sys

def default_filter(account, body):
	return open('default', 'ab')

def unknown_list_filter(account, body):
	if re.search("^X-List.*" ,body, re.M):
		return open('unknown_list', 'ab')

def y_groups_filter(account, body):
	if re.search("^(To)|(Cc)|(From):.*@.*" + \
		"(groups.yahoo.com)|(yahoogroups.com)", 
		body, re.M|re.I):
		return open('y_groups', 'ab')

def to_cs_filter(account, body):
	if re.search("^(To)|(Cc)|(From):.*@crestsource.com",
		body, re.M|re.I):
		return open('cs', 'ab')

def dump_attachments_filter(account, body):
	if re.search("multipart/mixed", body, re.M):
		return getmail.null_file() #dump it
		
def dump_big_mails_filter(account, body):
	if len(body) > 100000:
		return getmail.null_file()

def chain(account, body, chain):
	for link in chain:
		r = link(account, body)
		if r:
			return r

def chain1(account, body):
	return chain(account, body, [to_cs_filter, 
	unknown_list_filter, default_filter])

def chain2(account, body):
	return chain(account, body, [dump_big_mails_filter,
	y_groups_filter, dump_attachments_filter, 
	default_filter])

accounts = [
	('mail1.insom.me.uk', 'insom', 'pass', chain1),
	('mail2.insom.me.uk', 'simon', 'pass', chain2),
	('mail3.insom.me.uk', 'mison', 'pass', 'mail3'),
]

getmail.check(accounts, out=sys.stdout)

The above code is untested, but should represent a Chain of Responsibility pattern.

Caveats: This is a fragile program. Dropped TCP connections will throw fatal exceptions, as will unopenable files or full filesystems. That said, it won't DELE until it RETRs, so you shouldn't lose mail.

Also, and this is a big point, it reads the whole mail into memory - twice. If you are in the habit of receiving 50Mb 2-page Word documents then this may not be for you.