james-server-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kevin Jackson" <foamd...@gmail.com>
Subject Re: Accessing BayesianAnalysisFeeder through Java
Date Wed, 02 May 2007 01:34:41 GMT

> You have to use the JDBCBayesianAnalyzer class.
> It is abstract because you have to implement the logging method.
> This is done usually via anonymous classes:
> /**
>  * The JDBCBayesianAnalyzer class that does all the work.
>  */
> private JDBCBayesianAnalyzer analyzer = new JDBCBayesianAnalyzer() {
>     protected void delegatedLog(String logString) {
>         log("BayesianAnalysisFeeder: " + logString);
>     }
> };

Ok, this seems reasonable.

> If you look at BayesianAnalysisFeeder mailet sources you will see
> exactly how it instantiate the analyzer and how it calls it.

Ok, I will check out that code.

> You have to init the sql queries (there is an xml file declaring how the
> needed sql queries look like), then you can call
>                  analyzer.addSpam()+analyzer.updateSpamTokens() or
> analyzer.addHam()+analyzer.updateHamTokens();

Based on this, my understanding is that if I write the code using the
JDBCBayesianAnalyzer, then basically I can communicate between my
webapp (mail UI) and James via the database tables.  If this is the
case then that is the perfect way for me to proceed.  I had thought
that I would have to create a new email message and set the header to
X-isSpamProbability=1.00 then send to some random address with the
spam as an attachment, but this way is *much* easier.

Thanks for your help,

To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org

View raw message