SpamFilter
From Katrina Help Info
return to Spam Patrol
Introduction
Monday 5/8/2005
The SpamFilter project has been discussed here:
And is maintained here:
I need help from an experienced PHP programmer to modify the Spam Blacklist Extension. I have identified 3 layers for protection: (a) CSS markup patterns that are used by spammers; (b) URL related keyword patterns that are blocked conditionally and with adjustable threshholds; and (c) URL patterns which are blocked unconditionally. The three black lists would be maintained as wiki pages with restricted access. A user front end for tuning the layer logic would be a bonus.
Specificity increases from (a) to (b) to (c). If layers (a) or (b) block a potential spammer, layer (c) does not need an entry for that specific spammer, thus saving wiki moderators from much busy work.
There is a hand edited version of the 3 layer solution being used at http://www.katrinahelp.info/
The 3 layers are combined into on PCRE compatible regex. The results are monitored here http://www.katrinahelp.info/wiki/index.php/Spam_Patrol
AnnaLissa Cruz wrote a SpamFilter Ruby upload script and Rudi Cilibrasi set up a cron job to update the regex at KatrinaHelp.info.
- The raw lists are available here:
- The regex is avaliable here:
The project could be enhanced to include remote updates from trusted wikis.
--jwalling 23:49, 8 May 2006 (CEST)
