POC Applying Bayesian filtering to XMPP MUCs
- Java 100%
|
Tad
9c619cf983
Make builds more reproducible
https://docs.gradle.org/current/userguide/working_with_files.html#sec:reproducible_archives Signed-off-by: Tad <tad@spotco.us> |
||
|---|---|---|
| gradle | Enable strict dependency verification | |
| src | Tiny tweak | |
| .gitignore | Checkin http://classifier4j.sourceforge.net | |
| .gitlab-ci.yml | CI & Packaging | |
| build.gradle | Make builds more reproducible | |
| gradle.properties | Enable strict dependency verification | |
| LICENSE | Initial commit | |
| README.md | Note the prebuilts | |
| XMPP-Filter.iml | Initial commit | |
Bayebot
Overview
A proof-of-concept XMPP bot that applies a Bayesian filter to messages in a MUC and reports detected matches.
Quick Start
- Populate Messages-Spam.txt with known spam messages
- Populate Messages-KnownGood.txt with known good messages
- Populate Rooms.txt with rooms you want to monitor
- Populate Account.txt with first line JID and second line password for the bot account
Prebuilts
TODO
- Retroactive learning from MAM based on flag users
- Ability to kick/ban repeat offenders
Credits
- Tigase Java XMPP Client Library (AGPL-3.0), https://github.com/tigase/jaxmpp
- Classifier4J (Apache-1.1, with permission granted for use under Apache-2.0), http://classifier4j.sourceforge.net