Remove/Delete Attachments from email server (IMAP)

Question

I have a rather large quantity of emails on my mail server (IMAP) which contain attachments (PDF's, etc.) These attachments waste space--my HDD space on server is limited by my hosting company--and I am trying to regain some of that space by removing the attachments from the emails. (These are redundant on the mail server as they have pretty much all been downloaded or uploaded locally.)

When I used Windows (a long time ago) there was an app called "IMAPSize" which was a "client-type" email handler which could be used to login to mail server and manipulate the messages, remove attachments, etc. But I have not found (so far) anything comparable in terms of ease of use and getting the job done, for Unix/Linux (although we might have much better and more robust solutions--I just don't know which.)

I am aware of munpack (have installed and used it locally) and ripmime but it seems to me that those programs are meant to handle messages when they are stored locally, Mbox or MailDir. I do not sync my email messages locally, preferring to leave them on our server (have so many emails, over 100k, in aggregate.)

Moreover, Munpack seems to be great at extracting attachments but I have not seen a way to just delete the attachments, without destroying the message itself.

I do have Horde, Roundcube and SquirrelMail available as web GUI's but, again, have not seen a way to delete only the attachments.

Additionally, I am looking for a solution which does not involve complex setup and configuration, etc. (I may be dreaming/asking for too much.)

I thought of trying this approach: use offlineimap to download and sync all of my emails locally, then possibly use munpack or ripmime to delete attachments (provided I can find a way to do that, as opposed to only extracting them) and then re-sync with server and hopefully, emails would be back on server without attachments but a) not sure if this approach would work and 2) again, I am not really looking to download tens of thousands of emails, locally....

I am sure there is a solution out there for me (perhaps some type of client-type GUI app which would let me manipulate emails on servers/delete attachments; or some other not too complex approach.)

Note: I do not use Thunderbird and have no desire to install or use it; and same for Wine. I try to keep my box very 'minimal' with light footprint apps, wherever possible.

Thanks for a pointer in the right direction.

Answer 1

I had the following bit of perl that strips all attachments from stdin and returns stdout which may be of help:

#!/usr/bin/perl -w
use strict; 

use Mail::Audit; 
use Mail::Audit::Attach qw(Attach); 

my $mail = Mail::Audit->new; 

my $attachments = $mail->attachments; 

foreach (@$attachments) 
{ 
      $_->remove; 
} 
$mail->print();

and than a simple loop over the files in your Maildir you want stripped of attachments for isntance:

for filename in <list> 
do 
  ./strip.pl < "$filename" > "$filename".lock && mv "$filename".lock "$filename" 
  rm "$filename".lock
done 

An interesting modification could be to first extract the attachments and store them separately first before deleting them from the mails:

foreach (@$attachments) 
{ 
      $_->save("/path/to/attachment/dir");
      $_->remove; 
}

Make a back-up first though ;)

Answer 2

I tried IMAPSize as well, but it is no longer supported, crashed sometimes, and left behind messages that it seems it should have deleted. There is also the Unattach app for GMail users, though I can't vouch for it.

To solve it for myself, I made a python script that may be of use: IMAP Size Reducer

It works on one specified folder per run. As long as you know how to run python scripts on the command line, how to log in to your server (URL, username, password), and the IMAP folder specification (e.g. "INBOX.Sent"), it should work with any IMAP folder. (Your ISP can help with those details if necessary.)

You can specify the "max desired message size" (i.e. "if the message is larger than this size, remove attachments until it is under this size") and "tolerable part size" (i.e. "if an individual message part is under this size, don't delete it, no matter the overall message size").

The following MIME types are considered candidates for part deletion:

Anything with Content-Disposition "attachment"
Content-Type image/*
Content-Type audio/*
Content-Type video/*
Content-Type music/*
Content-Type x-music/*
Content-Type application/*

It has not been rigorously tested on various email providers, but it worked well for me. See the web page for usage details.

Answer 3

You could use Thunderbird and the AttachmentExractor Addon for this. Simply add your mail account to Thunderbird and then start the AttachmentExtractor. It can remove the content of attachments while keeping the attachments filename in your mail.

Answer 4

Removing the attachments on the server is quite easy, they are only (text) files. - Allocate mails, having an attachment (usually the larger ones):

find -P /home/*/mail/*/*/* -mindepth 1 -maxdepth 1 -name "*=*" -size +2M -mtime '+367'

It works 100% well on WHM server with mail directory, not tested with mbox

• Loop over the files
• if grep -e '------=_' ${file}; # containing strings .... then
• split the emails on common criteria, eg.: csplit --prefix=split4mail2clean. -z ${file} /commonstring/
• allocate/find within the split files you strings, attachments will have some typical strings, delete those parts
• join the temporary split files left (those without the typical attachments) in reverse order to rebuild the original email
• reapply the modification timestamp
• do other things here, if needed

In order to succeed, you might need 3, 4 possible more IF conditions and delete if file contains this string ... Run this shell script recommended from a temporary folder, adjust sizes and other search criteria. Yes, I have a working script, tested on some 40000 mail files and it's leaving 0 attachments, within the specified search and delete criteria.