Having spent over 2 hours this morning cleaning up a load of spam in the forums on moodle.org, laboriously deleting threads one by one, I'm hoping this improvement can be implemented one day.

Another note about spam cleanup on moodle.org being very time-consuming without access to phpmyadmin or an improved spam cleaner tool.

One spammer added multiple spam comments to almost every entry in http://moodle.org/books. Unfortunately the comments report was no help at all in deleting them.

Raj are you able to have a look at this soon? It's becoming quite an issue on moodle.org.

We can chat about it anytime.

Sure Martin,
I will look into this

Looking at Spam Cleaner, it uses simple sql query on user-description and post-summary. I can add more search sql, but my only concern is server load by adding all these sql.
IMO:

1. We might want to improve this by parsing all post etc. and report spams, rather then searching for it.
2. Blacklisted words should be modifiable rather then static array.

Totally agree that parsing at post time is a better idea, but that's another tool.

This tool is a sledgehammer - the approach here is

1) Identify definite spammers, and
2) Zap everything we can think of for them

Helen said that it would good if we searched for (at least) comments, messages and forum posts to identify spammers.

If we could just print a number for each user with the number of matches that would be good.

However, Helen also said that she usually knows the spammers, and just needs a tool to delete all their stuff.

So it's probably a good idea to first make a new block for moodle.org that we can put on the user profile page, with a nuclear bomb button that deletes all the content this user ever produced and also rewrites their profile, changes their auth to nologin and so on. Easier than modifying something in core.

Helen can you spec this on a separate issue?

Thanks Martin, I've created MDLSITE-1619 and MDLSITE-1620.

There are two commits:

1. Added sql for comments, messages and forum posts.
2. Added user_id and number of matches for each section with link. Not sure if this is back-portable.
   Also, spam cleaner can be further cleaned and sql can be optimized. Probably in another bug/improvement.

Hi Raj,
This is a great improvement!
Here are a few suggestions:-

• post.subject (blog subject) should be searched as well not just post.summary IMO
• the way user->description is used in filter_user function seems very hacky to me, might be a better to use a separate variable instead
  title
• since we are doing this in fill_user_entry() when user->id is already set:-

  $SESSION->users_result[$user->id]->count[$spamin] += 1;
  

  $SESSION->users_result[$user->id]->description will have the value of first spam encountered always.

• filter_user function should also have an if condition for the case when the spam came is from user's profile.
• Does filter_user function really need $count variable passed?
• +1 to built a tool that can handle this during posting. There are a few proposed structure for the tool open for discussion in forums.
  Thanks

Thanks for the feedback Ankit,

I know user->description usage is a bit hacky, I tried not to deviate from current implementation. Also, as per count variable, it was kept for BC. Why do you think, if statement is needed in filter_user?

I will add post.subject in search. For rest of the point, can you please confirm, if I should consider changing the code. Only reason, I kept that is to make minimum changes and make it BC. Although second commit is quite a change in functionality and not sure, if it should be done with minimum changes (As I have done it currently.)

Hi Raj,

• As you can see in the attached pic we have "From blog post:" etc to specify where the keyword was found. But when its found in the description, it just displays the text without any header explaining where it was found, that is the reason I was asking another "if clause" to be included for this case.
• IMO, we should make the changes required, the $user->direction things definitely needs to be changed. If not in Stables, than at least in the master. But again that's just me
  Thanks
  Edit: the attached image somehow ended being way too small in size, open in a separate window to see the enlarged version

Thanks for the feedback Ankit,

I have removed my second commit, as count and showing all spams, can't be done without a design change. Will open another bug to handle it nicely. Can you please review the new patch.

Looks good to me!
Thanks

Thanks Ankit,
Pushing it for integration review.

Integrated... thanks!

BUT, this is clearly a dead-end. We cannot keep adding queries there like crazy in that way IMO. They all are full table scans + OR operations, really not sustainable IMO.

We should prospect other alternatives asap IMO.

Ciao

works as expected
Thanks

Sorry I've not had time to test this cool improvement, but just wondering whether anyone can confirm that if you search for the word "ugg" that posts, comments etc. containing the word "suggestion" are NOT listed in the search results.

Hi Helen,
Looking at the code, we use simple wild card match %keyword%, which means 'suggestion' will be included if you are searching for 'ugg' as far as I can predict.
Spam cleaner definitely needs a lot more major improvements.
Thanks

Whoever decided one week was worth 14 days had really one bad idea. Anyway, the nightmare is over, so thanks for your, once again, amazing contributions. Many, many thanks!

Now... disconnect, relax and enjoy the next days, yay!

Closing...ciao