A complete rewrite of the spam module. This is not functionally complete (most notably missing an administrative interface for spam content), but basic filtering is working.
Includes four spam filter plug-ins:
- duplicate filter: track if the same piece of content is posted multiple times
- bayesian filter: auto-learn whether or not content is spam
- URL filter: utilizes bayesian filter to auto-learn whether certain URLs are spam URLs
- node_age filter: a tunable filter that assumes the older a node is, the more likely a comment posted to it is spam
Each filter can be assigned a weight. For example, if the duplicate module is more effective, you can weight it's score 250%. If the node_age filter is less accurate, you can only weight it's score 50%. Filter types automatically set their suggested defaults.
New filters can be easily implemented thanks to the spamapi.
Includes three content-type plug-ins:
- comment: filters comments before they hit the database, as well as when they are inserted/updated
- node: filters nodes before they hit the database, as well as when they are inserted/updated
- user: minimal support for filtering users, not yet a completely support content-type
Support for new content-types is easily added thanks to the spamapi.