Bots noticeboard |
---|
This is a message board for coordinating and discussing bot-related issues on Wikipedia (also including other programs interacting with the MediaWiki software). Although this page is frequented mainly by bot owners, any user is welcome to leave a message or join the discussion here. If you want to report an issue or bug with a specific bot, follow the steps outlined in WP:BOTISSUE first. This not the place for requests for bot approvals or requesting that tasks be done by a bot. General questions about the MediaWiki software (such as the use of templates, etc.) should be asked at Wikipedia:Village pump (technical). |
Bot-related archives (v·t·) |
---|
Bad bot edit
See Wikipedia:Bots/Noticeboard#Monkbot_18 and Wikipedia:Village_pump_(proposals)#Survey_(CS1), there is clearly not consensus this edit is appropriate. Given that Monkbot task has been halted, is it possible to stop Citation Bot from being used as a backdoor to make those non-consensus changes? Hog Farm Talk 22:23, 11 February 2021 (UTC)
- You're citing an RFC that opened a day ago asking a question that will result in Monkbot either being (re)confirmed or disallowed; there is no consensus at that RFC (yet). It has nothing to do with Citation bot's task or its consensus to run. Primefac (talk) 22:27, 11 February 2021 (UTC)
- @Primefac: My question is "if this is so controversial with monkbot that we are having an RFC, why is citation bot changing |accessdate to |access-date?". Shouldn't bother changing of |accessdate to |access-date by bots in general be put on hold until there is consensus as to if that is appropriate? Hog Farm Talk 22:40, 11 February 2021 (UTC)
- The controversy is whether we should have a bot making multiple millions of edits to remove a parameter that is deprecated. Citation bot makes many different types of edits as it goes about its tasks, the vast majority of which are non-cosmetic; if occasionally the only thing to fix is a deprecated parameter change like the one you listed, that's not much different than someone using AWB to make changes and occasionally making a "genfixes only" edit. Primefac (talk) 22:48, 11 February 2021 (UTC)
- @Primefac: My question is "if this is so controversial with monkbot that we are having an RFC, why is citation bot changing |accessdate to |access-date?". Shouldn't bother changing of |accessdate to |access-date by bots in general be put on hold until there is consensus as to if that is appropriate? Hog Farm Talk 22:40, 11 February 2021 (UTC)
- I agree with Primefac. It's one thing for a bot to be making many bad edits. It's another for the code to occasionally result in a cosmetic-only edit (it happens, especially in complex codebases). Many botops are inactive these days and I'm not a big fan of bugging them over an occasional cosmetic edit. The idea behind COSMETICBOT is watchlist spam, not that the edits themselves are bad (ergo, BOTPOL says
Keep in mind that reverting a cosmetic edit is also a cosmetic edit.
) - There's nothing BAG can do about this anyway. Revoking the approval, or blocking the bot, would not be a proportionate or appropriate response, so contacting the operator and making a request should be the first stage per WP:BOTISSUE. But I imagine such a suggestion for a code alteration would have more weight after the RfC is closed. However, if you believe that a vast number of Citation bot's edits are cosmetic then that would be a different matter, and we may be able to review that. ProcrastinatingReader (talk) 23:48, 11 February 2021 (UTC)
- Primefac and ProcrastinatingReader, I sampled the bot's edits during a half-hour period this evening (0100-0130 UTC). During this time by my count the bot made 122 edits that solely changed parameter hyphenation, representing just over half of the bot's total edits. I also looked for but did not locate a bot approval that covered these edits for this bot. Under these circumstances I agree with Hog Farm's assessment: the bot should not be making these edits. Nikkimaria (talk) 02:09, 13 February 2021 (UTC)
- Keep in mind also that the bot is user-driven, so if there's a common user among those edits they should be notified if there are problematic edits. Primefac (talk) 02:38, 13 February 2021 (UTC)
- @Abductive: ↑. Nikkimaria (talk) 02:39, 13 February 2021 (UTC)
- Citation Bot's BRFA history does confuse me a little bit, so I'm also curious which BRFA it's operating under. Not entirely obvious from User:Citation_bot#Bot_approval. Though I recognise many of its authorisations are from the stone age. The most recent BRFA is from 2011 which predates the modern COSMETICBOT provisions. It's probably something to clarify with the botop in the first instance, though perhaps some older BAG here may be able to offer some context too. ProcrastinatingReader (talk) 02:53, 13 February 2021 (UTC)
- CB's bot basically been converted to a user-activated tool ages ago, and its behavior more or less handled locally. There's plenty of oversight, just not BAG oversight specifically (althought plenty of BAG people watch that page). The main issue here is that the convertion of hyphenated parameters is a cosmetic task, and would be fine to do alongside other edits, but convertion hasn't been coded as cosmetic edits, so they are done on their own (which again, they shouldn't be). So simply follow WP:BOTISSUE, and file a bug report as outlined at User talk:Citation bot explaining the issue (i.e. the bot is making cosmetic edits), and things will get taken care of. There's no need to suspend or revoke anything at this point. Headbomb {t · c · p · b} 03:47, 13 February 2021 (UTC)
- Is there a list of all kinds of edits Cosmetic bot makes somewhere (that isn't the codebase)? I mean, if this bot has tasks not approved by virtue of being a tool, surely there should be a list of types of edits somewhere that it does, otherwise how does a person triggering the bot know what edits it's about to make? Compare to AWB genfixes, which I think are all documented at Wikipedia:AutoWikiBrowser/General fixes and some other pages. Re last sentence: not suggesting suspension/revocation; the operator is inactive and BAG doesn't have the resources to process approvals again anyway, mostly just curious on transparency. ProcrastinatingReader (talk) 15:14, 13 February 2021 (UTC)
- CB's bot basically been converted to a user-activated tool ages ago, and its behavior more or less handled locally. There's plenty of oversight, just not BAG oversight specifically (althought plenty of BAG people watch that page). The main issue here is that the convertion of hyphenated parameters is a cosmetic task, and would be fine to do alongside other edits, but convertion hasn't been coded as cosmetic edits, so they are done on their own (which again, they shouldn't be). So simply follow WP:BOTISSUE, and file a bug report as outlined at User talk:Citation bot explaining the issue (i.e. the bot is making cosmetic edits), and things will get taken care of. There's no need to suspend or revoke anything at this point. Headbomb {t · c · p · b} 03:47, 13 February 2021 (UTC)
- Keep in mind also that the bot is user-driven, so if there's a common user among those edits they should be notified if there are problematic edits. Primefac (talk) 02:38, 13 February 2021 (UTC)
- Primefac and ProcrastinatingReader, I sampled the bot's edits during a half-hour period this evening (0100-0130 UTC). During this time by my count the bot made 122 edits that solely changed parameter hyphenation, representing just over half of the bot's total edits. I also looked for but did not locate a bot approval that covered these edits for this bot. Under these circumstances I agree with Hog Farm's assessment: the bot should not be making these edits. Nikkimaria (talk) 02:09, 13 February 2021 (UTC)
Bot to report 3RR violations to AN/3?
I'm surprised there isn't a bot working at AN/3 to report 3RR violations, given AIV has DatBot and the like, and this is arguably a much simpler task. Did a bot like this exist in the past, or is there something I'm not thinking of that makes this a bad idea? --C o r t e x 💬talk 18:50, 16 February 2021 (UTC)
- We don't need such a User:JavertBot. If two people get into a 3RR or 4RR edit war and shift from the article to the talk page in the middle of it, the AN3 report would be counter productive. Yes you shouldn't revert war, but there's a lot of context that bots can't detect. Headbomb {t · c · p · b} 19:39, 16 February 2021 (UTC)
- It would need to ignore edits exempt from 3RR (which, just guessing, would make up a substantial percentage of naive hits for "a user making 3+ reverts to a single page in 24 hours"). This is unlikely to be a trivial task, and we would not want to fill the noticeboard with obvious vandalism reversions and the like. — The Earwig ⟨talk⟩ 19:39, 16 February 2021 (UTC)
- While a bot to make reports might be problematic, I still think that a bot that locks articles when there is an active edit war would be rather helpful, and save those dozen-edits-in-as-many-minutes issues with vandalism. Primefac (talk) 22:03, 16 February 2021 (UTC)
- Still a protection policy violation in many cases, where it’s just a single editor vandalising and being reverted or some such. Not reasonable to protect a whole article from future editors, for any period of time, due to a single editor I think. ProcrastinatingReader (talk) 22:16, 16 February 2021 (UTC)
- While a bot to make reports might be problematic, I still think that a bot that locks articles when there is an active edit war would be rather helpful, and save those dozen-edits-in-as-many-minutes issues with vandalism. Primefac (talk) 22:03, 16 February 2021 (UTC)
- See User:ProcBot/EW (and its page history) which tracks this stuff. Full of whack stuff, but also many false positives. I guess more admins could watchlist and monitor that page. ProcrastinatingReader (talk) 22:14, 16 February 2021 (UTC)
Extracting information from the Automatic taxobox template
ShortDescBot is currently working on organism articles, and to create its short descriptions needs to extract the taxonomic rank of the article subject. It can mostly do that quite well by naively parsing the text, checking for categories etc, but it has more trouble with articles that use Automatic taxobox, such as Lampshade spider, as the information isn't directly available within the wikitext. Is anyone able to help with some guidance on how I can access the rank (family
in that case) via Pywikibot/Python? MichaelMaggs (talk) 18:53, 16 February 2021 (UTC)
- If the article uses {{Automatic taxobox}}, extract the
|taxon=
parameter, then fetch Template:Taxonomy/taxon (Template:Taxonomy/Hypochilidae in this case), then pull the|rank=
from the template on that page. — The Earwig ⟨talk⟩ 19:35, 16 February 2021 (UTC)- The Earwig, perfect, thank you! MichaelMaggs (talk) 20:58, 16 February 2021 (UTC)
Page-blanking filter for bots
Hello, please see WP:EFR#Page-blanking filter for bots. I've asked that the page-blanking filter be reconfigured so that bots will be filtered from blanking pages (without any way to override the filter), thanks to the occasional error such as https://en.wikipedia.org/w/index.php?diff=1000987634. Your input would be greatly appreciated, of course. Nyttend (talk) 12:52, 17 February 2021 (UTC)
- You should do a thorough check that whatever your proposed filter does exactly doesn't conflict with the actions of properly functioning bots. The existing enabled filters I see that refer to "blanking" seem like they could false-positive on things like archiving a talk page that lacks other sections or many headers, AnomieBOT's TFDTemplateSubster removing a template from a talk page without much other content, bots that maintain lists of pages attempting to reflect that humans have cleaned up pages in the lists, and so on. Anomie⚔ 13:20, 17 February 2021 (UTC)
- I also agree that this needs careful thought. Perhaps only enforcing the filter in mainspace, and logging elsewhere, would be worthwhile? ƒirefly ( t · c ) 13:59, 17 February 2021 (UTC)
- @Nyttend: suggest you just duplicate the filter, apply it only to group=bot, put it in monitor mode and report back after it has been monitored for a bit. — xaosflux Talk 14:01, 17 February 2021 (UTC)
- AWB can also inadvertently blank pages. -- GreenC 15:32, 17 February 2021 (UTC)
- Sorry I wasn't clear — my goal was mainspace pages only. I know I said mainspace when I was writing up the request; I guess I deleted it by accident before I posted the request. Nyttend (talk) 15:34, 17 February 2021 (UTC)
- It seems a little absurd that we're reaching this point. It should be pretty straightforward to identify that a bot was accidentally given a blank page and abort. Or, check the HTTP status code is OK before proceeding, or use api.php instead of action=raw, etc. I don't know if the AbuseFilter condition limit is still an issue, but deferring this sanity check onto AF instead of just fixing the bot seems backwards. Legoktm (talk) 19:15, 17 February 2021 (UTC)
- @Legoktm: I'm 100% on board that this should be fixed bot-side. If the first conditions are bot and mainspace it shouldn't hurt the AF much though - really if any hits are coming to this filter they need to be followed up and addressed by the operators anyway (which should already be getting done for any known issues). — xaosflux Talk 19:19, 17 February 2021 (UTC)
- Agreed with Legoktm that going the edit filter route seems to be an over-reaction to what was essentially one poorly written bot task. – SD0001 (talk) 13:15, 19 February 2021 (UTC)
WP:PYWIKIBOT
Apologies if this isn't the right place—couldn't see a project talk page. Is this thing approved to run on en.wp? ——Serial 12:30, 24 February 2021 (UTC)
- All bots have to be approved by BAG, as per WP:BOTPOL - how those bots are written is ultimately irrelevant. Many, many bots on enwiki use Pywikibot, so yes, as long as you file a BRFA and get the task approved, using Pywikibot is fine. ƒirefly ( t · c ) 12:35, 24 February 2021 (UTC)
- Pretty much what firefly says. It doesn't really matter how a bot is coded or what frameworks it uses (AWB, pywikibot or from scratch) for the purposes of bot policy. Bots coded using pywikibot still have to go through approval like all others. There are also some pywikibot semi-automated scripts editors can use I think - those would fall under WP:ASSISTED like other types of semi-automated editing. ProcrastinatingReader (talk) 13:21, 24 February 2021 (UTC)