Re: Display Bayes tokens?

[Jonas Eckerman]

Peter Sabaini wrote:


I'd like to verify the tokens Bayes uses to classify;

[...]

Is this encoded in some way?


Yes.

If you use SQL for bayes you can use my plugin CollectTokens plugin to 
collect new tokens indexed by the encoded value used by the bayes 
system. That way you can look upp tokens and see what they were. Of 
course, you'll only be able to look up tokens that were learnet after 
you started using the plugin.


I've only tested the plugin with MySQL, but it shouldn't be hard to 
modify it to use another SQL system.


The plugin is available at


Regards
/Jonas
--
Jonas Eckerman
Fruktträdet & Förbundet Sveriges Dövblinda
http://www.fsdb.org/
http://www.frukt.org/
http://whatever.frukt.org/

Re: Display Bayes tokens?

[Hajdú Zoltán]

Hello,

The 'stored tokens' are the last 5 bytes of the SHA1 hashed 'real token'.
There is no (effective) way to decode the stored versions to their original 
format.

But you can verify that a given token does exists in the database (tkate the 
last 5 bytes of the SHA1 hashed token...).


Cheers,

Peter Sabaini írta:

Hi,

I'd like to verify the tokens Bayes uses to classify; consulting man sa-learn 
leads me to believe that "sa-learn --dump" or 
alternatively "sa-learn --backup" should do just that. 


However the output consists of lines which look like

t   3   28  1219966305  73e5bd1906
t   23  12  1223400311  672bc0c09a
t   20  92  1222528949  ef7de3e221

Is this encoded in some way? Is my Bayes' classifier mis-configured / 
mis-tuned?


Oh, this is with SpamAssassin version 3.2.5


TIA,
peter.

Display Bayes tokens?

[Peter Sabaini]

Hi,

I'd like to verify the tokens Bayes uses to classify; consulting man sa-learn 
leads me to believe that "sa-learn --dump" or 
alternatively "sa-learn --backup" should do just that. 

However the output consists of lines which look like

t   3   28  1219966305  73e5bd1906
t   23  12  1223400311  672bc0c09a
t   20  92  1222528949  ef7de3e221

Is this encoded in some way? Is my Bayes' classifier mis-configured / 
mis-tuned?

Oh, this is with SpamAssassin version 3.2.5


TIA,
peter.

-- 
http://sabaini.at