You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Matic Dolar <ma...@medium.si> on 2004/08/10 19:08:50 UTC
testing a whole mailbox
Hi!
I'm writing a study about spam and would like to do the following:
- run spamassassin against a whole mailbox file (presumably quite large one) full of mixed messages - both spam and ham (as well as a significant amount of viruses (albeit removed by antivirus-scanner).
- analyze the results I get - amount of spam received, amount of spam caught, etc ...
What command should I use to run spamassassin?
All I was able to achieve so far was to run it against a mailbox (plain text file) containing a single message. I want to make it analyze all the messages in the file, one by one.
Any ideas on how I could make this work or perhaps other suggestions?
Everything would be greately appreciated!
Tnx!
Matic
Re: testing a whole mailbox
Posted by Mike Burger <mb...@bubbanfriends.org>.
I don't think that there's a flag to just run against a mailbox. If I
recall, correctly, you can just use "spamassassin < /path/to/mailbox" and
it should read each message in that box, individually.
sa-learn, on the other hand, needs the --mbox flag for bayes training.
On Tue, 10 Aug 2004, Matic Dolar wrote:
> Hi!
>
> I'm writing a study about spam and would like to do the following:
> - run spamassassin against a whole mailbox file (presumably quite large one) full of mixed messages - both spam and ham (as well as a significant amount of viruses (albeit removed by antivirus-scanner).
> - analyze the results I get - amount of spam received, amount of spam caught, etc ...
>
> What command should I use to run spamassassin?
> All I was able to achieve so far was to run it against a mailbox (plain text file) containing a single message. I want to make it analyze all the messages in the file, one by one.
>
> Any ideas on how I could make this work or perhaps other suggestions?
> Everything would be greately appreciated!
> Tnx!
>
> Matic
--
Mike Burger
http://www.bubbanfriends.org
Visit the Dog Pound II BBS
telnet://dogpound2.citadel.org or http://dogpound2.citadel.org
To be notified of updates to the web site, visit
http://www.bubbanfriends.org/mailman/listinfo/site-update, or send a
message to:
site-update-request@bubbanfriends.org
with a message of:
subscribe
Re: testing a whole mailbox
Posted by Robert Menschel <Ro...@Menschel.net>.
Hello Matic,
Tuesday, August 10, 2004, 10:08:50 AM, you wrote:
MD> Hi!
MD> I'm writing a study about spam and would like to do the following:
MD> - run spamassassin against a whole mailbox file (presumably
MD> quite large one) full of mixed messages - both spam and ham (as well
MD> as a significant amount of viruses (albeit removed by
MD> antivirus-scanner).
MD> - analyze the results I get - amount of spam received, amount of spam caught, etc ...
MD> What command should I use to run spamassassin?
What you want is a mass-check run.
http://wiki.apache.org/spamassassin/MassCheck?action=highlight&value=mass-check
Bob Menschel
Re: testing a whole mailbox
Posted by William Stearns <ws...@pobox.com>.
Good day, all,
On Tue, 10 Aug 2004, LuKreme wrote:
> On 10 Aug 2004, at 11:08, Matic Dolar wrote:
> > - run spamassassin against a whole mailbox file (presumably quite
> > large one) full of mixed messages
>
> formail -s procmail -m sa.check < THE.MBOX
There's an alternate form that will process each of the individual
messages in the mailbox (calculating and adding the SpamAssassin score
block) without requiring procmail or a procmail config file:
formail -f -ds /usr/bin/spamc <origfolder >folder.reprocessed
This assumes you have spamd running. If you don't, I believe you
can simply substitute "spamassassin" for "spamc", although this will take
longer per message because of the increased startup time.
Note that LuKreme's approach with procmail will sort the mail into
spam and ham folders, while the above will just place all messages in
folder.reprocessed. Depends on what you want.
Cheers,
- Bill
---------------------------------------------------------------------------
bash$ :(){:|:&};:
-- Michal Zalewski
(Courtesy of George Bakos. _Don't_ run this.)
--------------------------------------------------------------------------
William Stearns (wstearns@pobox.com). Mason, Buildkernel, freedups, p0f,
rsync-backup, ssh-keyinstall, dns-check, more at: http://www.stearns.org
--------------------------------------------------------------------------
Re: testing a whole mailbox
Posted by Matic Dolar <ma...@medium.si>.
Thanks!!! This is exactly something I was looking for!
I'd just like to ask you one favor, if it isn't too much of a bother ...
Could you please explain what the lines in the script do (at least
approximatley), so I can play around with it?
Or could you at least point me in the right direction?
Thank you so much!
Matic
----- Original Message -----
From: "LuKreme" <kr...@kreme.com>
To: <sp...@incubator.apache.org>
Sent: Tuesday, August 10, 2004 9:19 PM
Subject: Re: testing a whole mailbox
> On 10 Aug 2004, at 11:08, Matic Dolar wrote:
> > - run spamassassin against a whole mailbox file (presumably quite
> > large one) full of mixed messages
>
> formail -s procmail -m sa.check < THE.MBOX
>
> sa.check:
>
> LOGFILE=the.mbox.log
> LOGABSTRACT=NO
> NL="
> "
> LONGDATE=`date "+%y-%b-%d %I:%m:%S %z"`
> TRAP='echo ""; formail -l "$LASTFOLDER" ;echo " Spam: $SPAM
> Deliverd: $LONGDATE'
>
> :0fw
> | spamassassin
>
> :0
> * ^X-Spam-status:\/
> { SPAM=$MATCH }
>
> :0:
> * ^X-Spam-Status:(.*\<)?Yes
> SPAM.MBOX
>
> :0:
> HAM.MBOX
>
> > - analyze the results I get - amount of spam received, amount of spam
> > caught, etc ...
>
> The logfile will contain the basic details on every email grouped into
> a single block
>
>
> --
> Ah we're lonely, we're romantic / and the cider's laced with acid / and
> the Holy Spirit's crying, Where's the beef? / And the moon is swimming
> naked / and the summer night is fragrant / with a mighty expectation of
> relief
>
Re: testing a whole mailbox
Posted by LuKreme <kr...@kreme.com>.
On 10 Aug 2004, at 11:08, Matic Dolar wrote:
> - run spamassassin against a whole mailbox file (presumably quite
> large one) full of mixed messages
formail -s procmail -m sa.check < THE.MBOX
sa.check:
LOGFILE=the.mbox.log
LOGABSTRACT=NO
NL="
"
LONGDATE=`date "+%y-%b-%d %I:%m:%S %z"`
TRAP='echo ""; formail -l "$LASTFOLDER" ;echo " Spam: $SPAM
Deliverd: $LONGDATE'
:0fw
| spamassassin
:0
* ^X-Spam-status:\/
{ SPAM=$MATCH }
:0:
* ^X-Spam-Status:(.*\<)?Yes
SPAM.MBOX
:0:
HAM.MBOX
> - analyze the results I get - amount of spam received, amount of spam
> caught, etc ...
The logfile will contain the basic details on every email grouped into
a single block
--
Ah we're lonely, we're romantic / and the cider's laced with acid / and
the Holy Spirit's crying, Where's the beef? / And the moon is swimming
naked / and the summer night is fragrant / with a mighty expectation of
relief
Re: testing a whole mailbox
Posted by Ryan Thompson <ry...@sasknow.com>.
Matic Dolar wrote to spamassassin-users@incubator.apache.org:
> Hi!
>
> I'm writing a study about spam and would like to do the following: -
> run spamassassin against a whole mailbox file (presumably quite large
> one) full of mixed messages - both spam and ham (as well as a
> significant amount of viruses (albeit removed by antivirus-scanner).
> - analyze the results I get - amount of spam received, amount of spam
> caught, etc ...
If you really want to do this right, you should use the tools in the
"masses" subdirectory of the SA distribution. You'll want to split the
mailboxes up into separate spam and ham corpora, and run mass-check on
each. From there, the spam.log and ham.log you obtain should tell you
almost everything you need to know.
http://wiki.apache.org/spamassassin/MassCheck
The masses stuff is designed with this purpose in mind. Although it is
tilted more toward rule analysis and scoring, those results are a
superset of most other results you might want to obtain, such as average
ham/spam scores, overall tagging accuracy, etc.
- Ryan
--
Ryan Thompson <ry...@sasknow.com>
SaskNow Technologies - http://www.sasknow.com
901-1st Avenue North - Saskatoon, SK - S7K 1Y4
Tel: 306-664-3600 Fax: 306-244-7037 Saskatoon
Toll-Free: 877-727-5669 (877-SASKNOW) North America
Re: testing a whole mailbox
Posted by Andy Jezierski <aj...@stepan.com>.
"Matic Dolar" <ma...@medium.si> wrote on 08/10/2004 12:08:50 PM:
> Hi!
>
> I'm writing a study about spam and would like to do the following:
> - run spamassassin against a whole mailbox file (presumably quite
> large one) full of mixed messages - both spam and ham (as well as a
> significant amount of viruses (albeit removed by antivirus-scanner).
> - analyze the results I get - amount of spam received, amount of
> spam caught, etc ...
>
> What command should I use to run spamassassin?
> All I was able to achieve so far was to run it against a mailbox
> (plain text file) containing a single message. I want to make it
> analyze all the messages in the file, one by one.
>
> Any ideas on how I could make this work or perhaps other suggestions?
> Everything would be greately appreciated!
> Tnx!
>
> Matic
Not sure 2.6x has that option, do a man spamassassin to be sure.
In 3.0 it's spamassassin --mbox filename
Andy