Gnutella Forums

Gnutella Forums (https://www.gnutellaforums.com/)
-   Feature Requests & Bug Reports (https://www.gnutellaforums.com/feature-requests-bug-reports/)
-   -   [Feature Request] Updated Default filters + auto-update (https://www.gnutellaforums.com/feature-requests-bug-reports/82011-feature-request-updated-default-filters-auto-update.html)

Klaus_1250 February 29th, 2008 06:05 AM

[Feature Request] Updated Default filters + auto-update
 
Would be nice to see some updated Default filters.

The adult-list if far from complete, and I think, should be split in to three or four categories. (e.g. "normal", illegal (<18 and such), perverse/extreme, ...). I'm pretty sure there are normal users who search for adult material on Gnutella but do not appreciate seeing the latter two categories I mentioned.

The Spam-filters could also use an update. They block some spam, but certainly not all. Will a few tweaks and updates, it could block much more spam.

Even better would be if the filters could be auto-updated (http or though gnutella), which would provide a much more timely and effective anti-spam mechanism.

arne_bab March 1st, 2008 07:04 AM

Would you like to help us, refine the filters?

Currently you can look at them in the following file: SourceForge.net Repository - [phex] View of /phex/trunk/src/main/java/phex/rules/DefaultSearchFilterRules.java

They work in the following way:

# Create the adult rule
Rule adultRule = new Rule();

# the condition when the filter gets active
FilenameCondition adultcondition = new FilenameCondition( );
adultcondition.addTerm( "blah" ).addTerm( "foo" )

-> activate when the search contains any one (or several) of the terms "Blah" and "Foo".

# Add the condition to the rule
adultRule.addCondition( adultcondition );

# What should happen: Don't show it in searches.
adultRule.addConsequence( FilterFromSearchConsequence.INSTANCE );

# Should it be active all the time?
adultRule.setPermanentlyEnabled( false );

# is it a default rule of Phex (can't be deleted)?
adultRule.setDefaultRule( true );

# Now assign it to the name.
ADULT_FILTER_RULE = adultRule;


If you want to contribute a default rule, you can just create it in phex and then give us the parameters, so we can add it.
You can see the terms we filter on for adult material in the link above. If you think, some should be changed, just tell us which :)

Alternatively you can naturally also post a patch...

Best wishes,
Arne

Klaus_1250 March 1st, 2008 11:29 AM

Sure, np. Will take a bit since I need to do some serious searching to get it a bit complete.

Btw here are two sources with adult keyword for Gnutella:
http://www.gnutellaforums.com/limewi...ds-filter.html
SourceForge.net Repository - [shareaza] View of /trunk/setup/filter/default.dat

arne_bab March 2nd, 2008 05:16 AM

Great!

Could you test, how well the filters in the links work when implemented as Phex rules?

Klaus_1250 March 2nd, 2008 11:29 AM

Will do, but it is going to take some time. Need to check, verify and see if there are not too many FP's given the limited filtering capabilities.

BTW: What's up with all those Limewire 4.12.8:6346 leafs running at datacenters?

arne_bab March 4th, 2008 12:14 AM

I'd assume the LW leafs might be spammers, but I'm not sure with that.

Do they offer good files?

Klaus_1250 March 4th, 2008 03:11 AM

Not that I can see (refuse browsing, report 1.5GB), but there isn't really a good way of checking it.

They connect quite aggressively e.g. if I remove them, they (or neighbouring nodes within the same range, same Vendor-string, same port) connect back in 10 seconds. They send/receive a lot of data compared to any other leaf (more in line of what you expect from an very active UP) and come from 3 or 4 neighbouring /24-blocks. Without blocking them, they always seem to find me.

On G2, I see similar behaviour (e.g. nodes in datacenters with /24 blocks at their disposal) and they do spam.

It's quite difficult to extract spammy leafs/UP with the current range of options available in Phex. E.g. if I open up the search monitor, I see dozens of what appear to be non-legitimate search queries, but there aren't any filtering options (UPs/Leafs/TTL/Hops) or ways to extract the data. I'm also not sure how to get a good glimpse of results, as I can only see results by a filter. I tried to set only one UP connection (while being an UP) for a better/easier view but this only seemed to work once.

arne_bab March 5th, 2008 01:14 AM

Do you have these blocks avaible?

You could test, if the behaviour improves when you block them in the security tab.

Blocking them in the security tab (which means banning them) drops all packets which contain their IP address, so they can't connect, can't send results, etc. (all as far as I know).

arne_bab March 7th, 2008 04:48 AM

1 Attachment(s)
I compiled the Blocks from Klaus into a blocklist file (attached).


All times are GMT -7. The time now is 01:11 AM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
SEO by vBSEO 3.6.0 ©2011, Crawlability, Inc.

Copyright © 2020 Gnutella Forums.
All Rights Reserved.