Extract only the body part of incoming emails using bash

Question

I use offlineimap to fetch the mails into a Maildir folder.

I want to automatically parse all new incoming emails in a Maildir folder and send only the "from", "subject" and "body" as an instant message somewhere else.

So I try to process all mails with

MPATH=~/Mail 

if [ -n "$(ls "$MPATH/INBOX/new/")" ]; then 
    for f in "$MPATH/INBOX/new/"*; do  
        SUB="$(cat "$f"|grep '^Subject' | head -n1 | sed "s/Subject: //g")"                                                                                       
        FROM="$(cat "$f" | grep '^From' | head -n1 | head -n 1|sed "s/From: //g")"                                                                                
        BODY="$(cat "$f"|sed -e '1,/Content-Transfer-Encoding/d')"
        MESS="$FROM: $SUB$BODY"

        echo $f 
        echo "$MESS" 
        mv "$f" "$MPATH/INBOX/cur/" 
    done 
fi

This already works fine for some simple emails, but how do I get rid of everything that is not the plain body, like signatures, attachements,...?

Dyann

Use a tool that understands mail instead of hacking it up with simple string processing? – Affenpinscher

I would, if there is a simple tool for the bash commandline? – Dyann

What are you actually trying to do here? "Read" your email? – Affenpinscher

Find an MDA you can configure to run commands on email as it gets delivered? – Affenpinscher

I use offlineimap to fetch the mails into a Maildir folder at the moment – Dyann

Check out formail to extract email headers – Rici

@glennjackman: How do you install formail? it is not in the debian repositories – Dyann

packages.debian.org/… – Rici

So i have to install the whole procmail? Can't i just get the formail bin somehow? – Dyann

Any luck with that? I'm looking for similar thing. – Squall

There's so very much wrong here. Start with mywiki.wooledge.org/ParsingLs – Honey

...and then why in the world would you cat a file over and over, instead of just reading it in one pass and parsing out the parts you want? – Honey

(...though if we needed MIME support, personally, I'd probably use Python's standard-library tooling rather than trying to reinvent the wheel). – Honey

I would be happy, of you put this in an answer with a solution to use – Dyann

Answer 1

As suggested on formail manpage, that worked for me:

cat mailfile | sed -e '1,/^$/ d' 

Calvo

Answer 2

As answered in the comments, formail from the procmail package seems to do the job nicely:

$ sudo apt-get install procmail

$ cat test.eml | formail -x To
 [email protected]

$ cat test.eml | formail -x Subject
 hello

$ cat test.eml | formail -x Content
 multipart/alternative; boundary="f403043eea78e8658a0554677278"

Credits: @glennjackman

Solangesolano