This is a repeat of the problem I was having earlier. After resetting
the databases (using Pickle AFAICT) and training on about 200 messages,
SpamBayes is lightning fast classifying messages. But over a couple of
weeks it gets slower and slower and slower, it's now taking 10+ seconds
per email to do its classification - and Python is maxing out at 100% of
cpu when doing it.
I have set it to not train on incoming email, not cache stuff, etc, to
no avail.
This is on a Mac, using Thunderbayes proxy.
Any suggestions?