strip-mbox-attachments.py

strip-mbox-attachments.py

#!/bin/env python

# Strips attachments from mbox, writes to new mbox.
# Does not load entire old mbox into memory first - useful for very large mboxes.
# Ref: https://ask.slashdot.org/story/11/12/04/1754257/ask-slashdot-handling-and-cleaning-up-a-large-personal-email-archive
# Ref: https://ask.slashdot.org/comments.pl?sid=2557794&cid=38260428

from __future__ import print_function
from mailbox import mbox, mboxMessage

print("Reading mailbox ...\n")
orig_mb = mbox('old_mbox_filename')
new_mb = mbox('new_mbox_filename')

print("Analyzing ...\n")
for key,msg in orig_mb.iteritems():
        print('.', end='')
        new_msg = mboxMessage()
        payload = msg.get_payload()
        if msg.is_multipart():
                payload = payload[0].get_payload()
                print('x', end='')
        for header in msg.keys():
                new_msg[header] = msg[header]
        new_msg.set_payload(payload)
        new_mb.add(new_msg)
new_mb.flush()