Wikidata talk:Abuse filter

From Wikidata
Jump to navigation Jump to search
SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 2 days. For the archive overview, see Wikidata talk:Abuse filter/Archive. The latest archive is located at Wikidata talk:Abuse filter/Archive/2024.

Occupation as P31[edit]

Could something be done to give (new) users the advise to use instance of (P31)human (Q5) instead of instance of (P31)your occupation here? This would avoid some returning constraint violations. Sjoerd de Bruin (talk) 14:41, 27 January 2016 (UTC)[reply]

Discussion
Given that filters only have access to added and current data, it could only work with a dedicated list of items which I'm afraid could be very expensive. Matěj Suchánek (talk) 15:07, 27 January 2016 (UTC)[reply]
Bummer. Sjoerd de Bruin (talk) 20:05, 15 February 2016 (UTC)[reply]

Block self-referencing first names[edit]

Common constraint violation, would be better to avoid this. Sjoerd de Bruin (talk) 20:05, 15 February 2016 (UTC)[reply]

Discussion


Tag changes to sourced statements[edit]

Changes to the value (not qualifiers, sources, etc.) of statements that have full sources (excluding "imported from") should be tagged. --Yair rand (talk) 19:52, 23 November 2015 (UTC)[reply]

Does anybody know a clear algorithm for this? Matěj Suchánek (talk) 18:05, 3 December 2015 (UTC)[reply]
Editor: IP; Property data type: number; reference: different from none or 'imported from'; edit: 0–9 or decimal point (but not a qualifier) ??? Is this an overview? --Chris.urs-o (talk) 06:56, 27 February 2017 (UTC)[reply]
Of course but try to explain it to an abuse filter (in particular, the reference part is problem). Matěj Suchánek (talk) 15:00, 28 February 2017 (UTC)[reply]

About: proposals#most important pages[edit]

Quote: "A not-confirmed user changes the string of the English label of Qid < 1,000,000 while the string occurs x-times (20, 30, 50) in the item."
Note: Not only the string gets vandalism, but label, date of death and date of birth as well. Regards --Chris.urs-o (talk) 12:09, 8 February 2016 (UTC)[reply]

Change of value of member of sports team (P54)[edit]

Is it possible to track changes of values to member of sports team (P54)? Most of those should be actually new statements or it is vandalism. Adding, changes or removal of qualifiers should be ignored. Sjoerd de Bruin (talk) 16:13, 20 June 2017 (UTC)[reply]

Mainspace links to Wiktionary[edit]

Mainspace links to Wiktionary should be disallowed by the abuse filter with some explanation. Sjoerd de Bruin (talk) 19:12, 27 September 2017 (UTC)[reply]

+1. --Yair rand (talk) 01:02, 28 September 2017 (UTC)[reply]
They are. Matěj Suchánek (talk) 14:27, 28 September 2017 (UTC)[reply]
It seems like people just ignore the message then. Can we disallow as well? Sjoerd de Bruin (talk) 15:07, 28 September 2017 (UTC)[reply]
Should we? I believe there are exceptions which shouldn't be blocked (like Wikimedia main page (Q5296)). Matěj Suchánek (talk) 15:22, 28 September 2017 (UTC)[reply]
I suspect that the main page may be the only exception, unless some wikis are using pseudo-namespaces. --Yair rand (talk) 00:11, 29 September 2017 (UTC)[reply]
This doesn't seem to work for links added via Special:NewItem. Sjoerd de Bruin (talk) 20:51, 5 November 2017 (UTC)[reply]
Yep :/ Matěj Suchánek (talk) 07:45, 6 November 2017 (UTC)[reply]
phab:T179810. Sjoerd de Bruin (talk) 08:07, 6 November 2017 (UTC)[reply]

Sexually transmitted infections[edit]

I think that bacterial STIs (chlamydia, gonorrhea, and syphilis) and viral STIs (genital herpes, HIV/AIDS, and genital warts) as english label for humans should be flaged as abuse. --Chris.urs-o (talk) 20:09, 11 February 2018 (UTC)[reply]
Which items are you referring to? How often does that happen? Couldn't we apply a protection? Matěj Suchánek (talk) 09:15, 12 February 2018 (UTC)[reply]
I was not watching Andrés Bello. It was Gonorrhea since november 17th. IP vandalism... --Chris.urs-o (talk) 16:56, 12 February 2018 (UTC)[reply]

Warning text for #102[edit]

Can we have a warning message for filter #102 such as "Lexemes cannot currently be used on items or properties due to a bug. See phab:T195611 and phab:T195615 for details."? The default "we detected that your edit is disruptive" message is not the most informative or friendly. Thanks. – Pizza1016 (talk | contribs) 06:07, 20 June 2018 (UTC)[reply]

✓ Done Matěj Suchánek (talk) 15:02, 20 June 2018 (UTC)[reply]

New abuse filter for adding interlanguage links[edit]

there're several hundreds uses of traditional interlanguage links (incomplete list), most of them should be interwiki link ([[:en:Page]]) instead. Probably we need an abusefilter to catch them (but do not catch them in user namespace, and warn only as we also have some cases like <nowiki>[[en:Page]]</nowiki>).--GZWDer (talk) 19:03, 13 August 2018 (UTC)[reply]

Apparently, interwiki links on Wikidata don't work and are equivalent to [[:en:...]]. Matěj Suchánek (talk) 11:52, 17 August 2018 (UTC)[reply]
@Matěj Suchánek: It does work (see sidebar of Wikidata:Property_proposal/Archive/28).--GZWDer (talk) 14:13, 23 August 2018 (UTC)[reply]
Sometimes they do not: en:Page. Need to investigate. Matěj Suchánek (talk) 10:12, 24 August 2018 (UTC)[reply]
@Matěj Suchánek: this is phab:T28085.--GZWDer (talk) 10:57, 24 August 2018 (UTC)[reply]
@Matěj Suchánek: I think this should be considered given Wikidata:Project_chat#fix.--GZWDer (talk) 19:46, 23 January 2020 (UTC)[reply]

Special:AbuseFilter/49[edit]

This filter should not report any additions of JORFSearch organization ID (P6413) or Image Archive, Herder Institute (P6482).--GZWDer (talk) 17:05, 11 February 2019 (UTC)[reply]

✓ Done Matěj Suchánek (talk) 18:49, 11 February 2019 (UTC)[reply]

Control and format characters[edit]

We probably need an abuse filter to track addition of control (Cc) and format (Cf) characters in labels, descriptions, aliases and values. Examples of Unicode format characters are soft hyphen, zero-width space and byte order mark. See also phab:T234136.--GZWDer (talk) 16:48, 29 September 2019 (UTC)[reply]

In addition, the replacement character (�) should also be tracked. But it might be more common and a separate filter may be needed.--GZWDer (talk) 16:50, 29 September 2019 (UTC)[reply]
Special:AbuseFilter/125 will track this. --Matěj Suchánek (talk) 08:58, 5 October 2019 (UTC)[reply]

Special:AbuseFilter/49[edit]

Please exclude prime factor (P5236).--GZWDer (talk) 05:59, 10 July 2021 (UTC)[reply]

✓ Done --Matěj Suchánek (talk) 16:35, 10 July 2021 (UTC)[reply]

Special:AbuseFilter/110[edit]

This filter is too broad - I was not able to fix the ISNI of Q887543 (UK Board of Trade) due to this. The ISNI of that item currently refers to the Swedish Board of Trade. ネイ (talk) 16:26, 19 July 2021 (UTC)[reply]

Abuse Filter 76 false positive when changing badges[edit]

{{Edit request}}

I tried adding intentional sitelink to redirect (Q70894304) to the Wikidata sitelink in Help:Wikidata (Q28925727), but got stopped from doing so by Special:AbuseFilter/76, because it doesn’t distinguish between sitelinks and badges:

— ExE Boss (talk) 13:20, 28 July 2021 (UTC)[reply]

Should be ✓ fixed. --Matěj Suchánek (talk) 18:43, 1 August 2021 (UTC)[reply]
@Matěj Suchánek: It’s still broken, see Special:AbuseLog/19714649.
Using summary rlike "\bwbsetsitelink-(add|set)(?!-badges)\b" will at least allow changing the badges using the API. — ExE Boss (talk) 23:00, 25 August 2021 (UTC)[reply]
✓ Done Feel free to test and report more cases. --Matěj Suchánek (talk) 09:02, 26 August 2021 (UTC)[reply]

Special:AbuseFilter/38[edit]

Since Q-numbers can now be 9 digits, this filter should use \bq\d{1,9}\b instead of \bq\d{1,8}\b. See these three edits on Wikidata Sandbox 2 (Q13406268). --Pokechu22 (talk) 18:34, 1 October 2021 (UTC)[reply]

✓ Done --Matěj Suchánek (talk) 09:16, 2 October 2021 (UTC)[reply]

Vandal removing sitelinks[edit]

A filter should be created to block removals of sitelinks from this cellular IP range. Thibaut (talk) 08:32, 6 June 2023 (UTC)[reply]

see Wikidata:Administrators'_noticeboard#Report_concerning_User:2001:44C8:4400:AB0A:D1E4:3908:308B:42E2 Estopedist1 (talk) 10:41, 6 June 2023 (UTC)[reply]
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata_talk:Abuse_filter&oldid=1909485645"