Discussion

Discussion

• I see the discussion is slow. In the optimistic spirit of "no news is good news," I'll add my two cents that, as the developer of this bot, I support permission for this task. Not only has it already run in the past within responsible bounds, I the developer already have a track record of running read-only tasks against the wiki community's open API for educational and research purposes that improve access to and interest in the wikimedia community. I'm of course open to input. Thank you for considering my request. Enfascination (talk) 23:32, 5 April 2026 (UTC)

  Since it's been seven days since the filing of the request: {{BAG assistance needed}}

  (Non-BAG member comment): To provide you at least some engagement on your thread: my understanding is that typically, read-only bots do not require BRFA approval (per this discussion). However, your description of the task suggests that WP:BOTPERF, including the database downloads, may be salient in your situation. Then again, perhaps a BAG member will have a different take here. Staraction (talk · contribs) 01:44, 7 April 2026 (UTC)

  These are the questions I have too: why are other solutions such as database downloads not sufficient to fetch the information you need, that you instead want to make up to 200K requests over a few days? Anomie⚔ 22:40, 7 April 2026 (UTC)

  Thank you @Staraction (and @Anomie). My answer, linked and elaborated below, is 1) that BRFA is being offered by WMF as a best practice for some very new (March/April 2026) rate limits on read-only bots, and 2) my pipeline is based on pre-Toolforge legacy code that, with the community's permission, I would like to be spared from rewriting. I understand if 200K is a lot, but it is also, by the 2001-2026 standards, not a lot. I appreciate your engagement and consideration. Enfascination (talk) 04:22, 8 April 2026 (UTC)

  I note that getting a bot flag would just let you get around some of the new automated rate limiting. It wouldn't stop 200K requests in a few days from getting the attention of humans at WMF who're looking for AI-crawler-like activity, and approval here may not make a difference to them. OTOH, it's possible that 200K requests is still small compared to what they're looking at, we don't really know. wikitech:Robot policy#Generally applicable rules suggests considering if other methods might be more appropriate, so we ask that. Anomie⚔ 22:29, 8 April 2026 (UTC)

  > approval here may not make a difference to them

  The policy is explicit that community approval is sufficient .

  But I understand that Toolforge is the more respectful use of WMF's finite resources. I will explore it before trying to lean more heavily on this community's forbearance. Thank you for engaging. Enfascination (talk) 03:28, 14 April 2026 (UTC)

• A bot flag isn't necessary for readonly tasks. We don't have exactly have a process either for such requests. If you are getting hit by rate limits when querying the API, you can run your data collection on Toolforge which is largely exempt from the new rate limits. Also pinging User:DKinzler (WMF) if they have any other information or advise around overcoming rate limit issues. – SD0001 (talk) 05:16, 7 April 2026 (UTC)

  Thank you. You correctly divined that the motivation for this request is that I was encountering the WMF's new 2026 read-only rate limits, just a few days old. That is probably a large part of why we don't exactly have a process for them. With these changes are new practices WMF is offering (authored in part by User:DKinzler, as you can see ). Given that they are so new, there is a chance I won't be the last to make this request, so it may be worth discussing whether the community will be open to adapting the BRFA process to consider read-only requests. You'll see that according to the table that, short of being a famous Wikipedian, my only options for quite high query rates are Toolforge and BRFA. In my case, this is legacy code from 2020, before Toolforge. It would be onerous to rewrite, in part because I inherited it from a departed student, and have been getting by on minor modifications for additional metadata as my coauthors need it. So I guess I'm asking if, as a public servant with a track record of serving the community responsibly, if I can be spared a rewrite. I enjoy bringing my background to Wikipedia, and I respect it a lot, but this is ultimately a side project, and I'm at risk of just giving up. If the norm is only to grant BRFA as an absolute last resort, then I understand a rejection. But being that BRFA is among the appropriate solutions being recc'd, I wonder if I could solicit the community to consider my request despite the fact there exist alternate ways I could have accomplished it if I had started today.

  I'm also comfortable doing additional due diligence as I win the community over on this borderline query. Thank you again. Enfascination (talk) 04:16, 8 April 2026 (UTC)

  The python code you linked above should be perfectly capable of being run on Toolforge. There's no need to rewrite it. Toolforge is a just a hosting platform, which supports Python (see wikitech:Help:Toolforge/Python) among other languages. I see some SQL files as well – ToolsDB can be used if the project needs its own database.

  That being said, if Toolforge really isn't possible without much effort, a bot flag can be considered. – SD0001 (talk) 05:57, 8 April 2026 (UTC)

  @SD0001 You're right: adapting to Toolforge seems straightforward. Thank you for the suggestion. I will follow the steps and see how it goes.

  > That being said, if Toolforge really isn't possible without much effort, a bot flag can be considered.

  Thank you for your flexibility. I won't come ask this community for a botflag again without having earned it with some due diligence on my end. Enfascination (talk) 03:31, 14 April 2026 (UTC)

Discussion

Should something similar be done with bold? (10,000 per this search) -- WOSlinker (talk) 21:46, 7 February 2026 (UTC)

Likely. It might also be worth requesting this be added to the genfixes for AWB so that when this run is over any new instances will be more likely to be picked up. Primefac (talk) 21:50, 7 February 2026 (UTC)

Yeah, I think it'd definitely be nice to do the same thing with erroneously bolded commas. I intentionally kept the query constrained to start off (ignoring any italicized terms with unusual characters, for instance), but it could be expanded after the initial run is over.

And yes, I agree it'd be nice to add this to the GENFIX set. Cheers, Sdkb ^talk 22:54, 7 February 2026 (UTC)

Are you not wanting to do bold? Primefac (talk) 17:48, 15 February 2026 (UTC)

I looked through the first 100 search results for the bold query. I found one niche edge case: On this page, bolding is used to delineate which parts of two passages match. Because manual line breaks are used, some bolded strings end with a comma. You could argue that this is a downstream effect of the article using poor syntax with manual line breaks, or that a passage like that should have been surrounded with {{as written}}. But because bolding is sometimes used for niche purposes like this, I think it's the slightest bit riskier to try to fix it than italics.

I'll defer to whatever the consensus is here about whether, given this, it's worthwhile to include it or not. Sdkb ^talk 17:44, 20 February 2026 (UTC)

This feels like something so minor that it would be best either ignored or done as part of AWB GENFIXES. I oppose this being done as the sole edit to a page. Thryduulf (talk) 14:28, 20 February 2026 (UTC)

It's certainly not the most earth-shattering change to a page, but it is an improvement, and it's clearly in compliance with WP:COSMETICBOT because it changes the output HTML of the page. It is something that I occasionally notice as a reader. Also, because it's an AWB bot, it can be run alongside GENFIXes, so often the comma fix will not be the only change the bot makes. Sdkb ^talk 17:20, 20 February 2026 (UTC)

I think we'll have to agree to disagree on whether the change is an improvement or neutral, and I have no objection to the change being made alongside changes that are unambiguously improvements, but minor changes like this should never be the sole change made by a bot. Thryduulf (talk) 18:40, 20 February 2026 (UTC)

On hold. There is opposition to the task, and with only the implication of consensus to run the task based on existing guidelines I would prefer to see a stronger consensus to specifically target this as a bot run. I know AWB releases updates less frequently than most countries change leadership, but that would be another route to go down to start whittling away at the list. Primefac (talk) 20:17, 8 March 2026 (UTC)

@Primefac, where would be an appropriate venue to get additional input on whether there is consensus to run this as a bot task? Thryduulf's view seems to be that WP:COSMETICBOT should be made stricter, and while I know that's a view some editors hold, presumably it's a minority given that editors have not found consensus to change the language of the bot policy. Sdkb ^talk 20:34, 8 March 2026 (UTC)

Either at the MOS talk or a Village Pump. I wouldn't necessarily say that it's a more strict ruling on COSMETICBOT given that it already says Minor edits are not usually considered cosmetic but still need consensus to be done by bots. Since this is a "barely visible" type of minor edit, I'd like to get at least some measure of support for making it; it's not like you're going to need an RFC, just enough to indicate that Thryduulf is in the minority when it comes to being concerned. Primefac (talk) 20:48, 8 March 2026 (UTC)