>> * There are experimental MySQL and PostgreSQL database backends,
>> which should have no problem with concurrent access. However, (a)
>> you would require an SQL server, and (b) I don't think these have
>> been tested with the Outlook plug-in.
>
> I can install MySQL - I've already done this on another server. Can
> anyone
> confirm whether or not this has ever been tried with the Outlook
> plugin? If
> not, how likely is this to work?
Somewhat likely.
> Where do I find the necessary code and
> information to allow me to try this?
Make sure you're using 1.1 (a1, a2, or a3 - it shouldn't matter).
Open the bayescustomize.ini file in your data directory (the Advanced
tab of the Manager dialog can open that for you). Add something like
this to the file:
persistent_use_database:mysql
persistent_storage_file:host=myserver user=myusername pass=mypassword
dbname=databasename
The database should already exist, but SpamBayes will create the
needed table when it's first used.
>> * If the messages are also in a shared location (Exchange or IMAP
>> server), then if you didn't use the Spam/Not Spam buttons but just
>> moved the messages, then one computer could do all the classification
>> and training.
>
> Not quite sure what you mean. The messages are in a shared location
> (on an
> Exchange server), but surely if I just move the messages to the
> spam folder
> by dragging them in Outlook, I won't get any incremental training.
By default, training when messages are moved is turned on (see the
Training tab of the SpamBayes Manager dialog). This (only run
SpamBayes on one machine) would be, by far, the simplest solution,
and should work without problem.
>> * You can use ZEO with 1.1a2 (or 1.1a3 if you run from source).
>> Like the SQL options, this requires setting up a (ZEO) server, but
>> unlike the SQL options, this has been tested (although not
>> extensively).
>
> Not quite sure what ZEO is - is this another database, or something
> different? My first inclination is to try to make it work with MySQL.
ZEO allows multiple processes, machines, and locations to work
together as one transactional object system. It basically allows
shared access to a database (typically ZODB). It's developed by Zope
Corp.
>> Alternatively, you could keep the databases separate, and consolidate
>> them periodically (e.g. daily).
>
> How would I do this consolidation?
The simplest way would be to use sb_dbexpimp.py (there isn't a binary
version of this, but you'd just need Python & the SpamBayes source).
For example (from the documentation):
Create a new DBM database (newbayes.db) from two
DBM databases (abayes.db, bbayes.db)
sb_dbexpimp -e -d abayes.db -f abayes.export
sb_dbexpimp -e -d bbayes.db -f bbayes.export
sb_dbexpimp -i -d newbayes.db -f abayes.export
sb_dbexpimp -i -m -d newbayes.db -f bbayes.export
=Tony.Meyer
--
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes.
http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this.
_______________________________________________
[email protected]
http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html
