Swpb bot (talk · contribs)

Operator: Swpb (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought:

Migration of structured data from the deprecated qualifier of (DEPRECATED) (P642) to replacement properties standardized by use case, following the process specified on Wikidata:WikiProject Deprecate P642/Migration process.

I have used a version of this script on Wikidata to migrate over 800,000 qualifier claims over the last 12 months, with very few mistakes reported, and those quickly corrected. I began using it on Commons two days ago, not realizing this ran afoul of bot policy, for which I apologize. I am creating this request immediately after being told I needed to, and I will not run the script on Commons again except under the requested bot account.

Automatic or manually assisted: Automatic, supervised

Edit type (e.g. Continuous, daily, one time run): Series of manually-started runs; script completes a set of edits taken from a SPARQL-derived job file passed to it, then stops.

Maximum edit rate (e.g. edits per minute): wikibase-cli uses a maxlag parameter to throttle batch editing when server load is high. This script currently uses the default of 5 seconds, but that can be set to a nicer value if needed. Rate with no server lag appears to be around 100/minute.

Bot flag requested: (Y/N): Y

Programming language(s): PowerShell script using wikibase-cli and jq

Swpb (talk) 16:28, 13 June 2025 (UTC)

Discussion

• Looks good to me. --Krd 07:49, 14 June 2025 (UTC)

Lingua Libre Bot (talk · contribs)

Operator: Yug (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought: 1,400,000 files to EDIT (SDC claims creation/update) in the Category:Lingua Libre pronunciation.

Automatic or manually assisted: Automatic supervised.

Edit type (e.g. Continuous, daily, one time run): one time run.

Maximum edit rate (e.g. edits per minute): 500 / min. (under urgency clause as Lingua Libre will be offline while migrating)

Bot flag requested: (Y/N): N (already have it)

Programming language(s): Python.

Yug (talk) 13:29, 13 June 2025 (UTC)

Discussion

Lingua Libre Bot is already an approved bot for regular updates.

This time, Wikimedia France is revamping the mass recording tool Lingualibre.org to a new Python Django VueJS codebase. This implies updating the Commons Structured Data of 1.4 million files in Category:Lingua Libre pronunciation & subcategories, over a single week ideally. The script is coded and working as expected, as visible in the tests below:

We therefore ask for approval to a new, massive, single shoot usage. We also request higher speed since the migration requires to take Lingualibre.org offline and the normal 5 secs bot speed would requires 70 days. Bot speed of 500 edits/min would allow the migration within 5 days. -- Yug (talk) 13:29, 13 June 2025 (UTC)

Could bot set all properties in single edit? --EugeneZelenko (talk) 13:56, 13 June 2025 (UTC)

GraphBot (talk · contribs)

Operator: GalStar (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought:

• Generate the necessary .chart and .tab files for graphs that use the deprecated Graph Extension
  • At the moment I feed the bot the article where the graphs need updating. It needs manual interval to specify the title of the graph.

Automatic or manually assisted: I manually specify what wikipedia articles to extract graphs from.

Edit type (e.g. Continuous, daily, one time run): Continuous

Maximum edit rate (e.g. edits per minute): anything it needs to be, since it's manually assisted the rate is limited

Bot flag requested: (Y/N): Y

Programming language(s): Rust, Python

GalStar (talk) 06:45, 3 June 2025 (UTC)

Discussion

• Interesting, this seems very useful. Is this done for all articles that use the graphs or how is generation … manually triggered? Also see graphDataImport. --Prototyperspective (talk) 09:54, 3 June 2025 (UTC)

  So my idea here is to slowly ramp up the amount that is automated. Occasionally graphs are broken and I need to double check everything (I believe most of the bugs have been fixed). Also the code itself is prone to panicking (i.e. aborting) if anything is out of line. At the moment the interface for me is the following:

  *:> cargo run -- "[Name of Article]" *:[bunch of diagnostics] *:Enter name of first graph > [user input] *:[more diagnostics] *:Enter name of nth graph > [more user input] *:

  And as each graph name is inputted the appropriate chart/tab files are created. Currently the main issue is that wikipedia articles often do not provide graph titles (see https://en.wikipedia.org/wiki/Demographics_of_Albania for examples).

  I have been considering having a talk-page template that could somehow mark the names of the graphs, but I need to flesh that idea out a little, and maybe ask the en-wikipedia folks for suggestions on that front.

  I have graph data import enabled as an user script, but it doesn't work for all cases (from the few times I used it, it made malformed graphs and I needed to use graphbot to generate the correct .tab file content).

  Cheers, GalStar (talk) 17:13, 3 June 2025 (UTC)

  Actually this is no longer the case. en:User:GraphBot has some details, but essentially it is fully automated now. GalStar (talk) 02:53, 6 June 2025 (UTC)

I've created some test edits for review GalStar (talk) 18:55, 3 June 2025 (UTC)

• relevant discussion on enwiki: en:Wikipedia:Bots/Requests for approval/GraphBot. —usernamekiran (talk) 21:10, 7 June 2025 (UTC)

• The wikipedia side of the bot has now been approved for a trial of 50 edits. If there are no objections, I will perform the trail, which will result in the creation of 100 files. I can review each file afterwards if needed. GalStar (talk) 17:36, 8 June 2025 (UTC)

  Please do. Krd 16:44, 12 June 2025 (UTC)

BoreBot (talk · contribs)

Operator: Borealex (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought:

• download hundreds of files using a pywikibot with pre-generated description templates;
• adding statement "captured with" (P4082) to structured data (code example).

Automatic or manually assisted: Automatic/supervised

Edit type (e.g. Continuous, daily, one time run): daily

Maximum edit rate (e.g. edits per minute): ~6 edits per minute

Bot flag requested: (Y/N): N

Programming language(s): Python (using Pywikibot)

Borealex (talk) 21:34, 29 April 2025 (UTC)

Discussion

Little clarification about adding structured data: the bot will not randomly sort through the metadata of all files, but will process files from ready-made lists received via the SQL-request (from Quarry for example). Test run: 1 2 3 4 5 6 7 8

What is that useful for? Is there any community consensus that this should be done a large scale? --Krd 13:36, 5 May 2025 (UTC)

Hi! I definitely know of at least one practical use for captured with added statements — the ability to search for examples of photos from various cameras, smartphones, camera lenses, scanners etc. This may not be a native user-friendly search functionality yet, but it's already working to some extent. The ability to make such requests useful, for example, when comparing devices before purchase. Some photohosting sites, such as Flickr, have similar functionality. And as my first step, I would like to contribute to filling the Wikimedia Commons with such information.

Answering your second question, I assumed that bots already exist that, among other things, perform the function of adding statements from EXIF metadata (for example), and if I understand correctly, there is a consensus on adding structured data by bots. The original idea was to create and fill categories, but I abandoned it in favor of structured data, based on the discussion of the BotAdventures discussion, where structured data is also given priority in this case. I would also add that the main difference between my bot and others will be the manual compilation of file lists according to something like this flow — I add a new device on Wikidata or add information about the Camera Model, create an SQL query and run the bot based on it. Borealex (talk) 20:24, 5 May 2025 (UTC)

I think structured data shall be set by bot only if they are useful, and I'm not convinced that it is relevant in more that a few dozen cases which camera a photo has been taken with. I think there should be specific community consensus before this can be approved to be applied to all files. Krd 07:13, 11 May 2025 (UTC)

Got it. Could you tell me please where would it be more relevant and appropriate to discuss this? Structured data Discussion page or Village pump? Borealex (talk) 18:23, 11 May 2025 (UTC)

I'm not sure, perhaps chose one of it and leave a link to it on the other? Krd 05:03, 17 May 2025 (UTC)

Started a discussion.

Commons:Village pump#Adding 'captured with' statement by bot

Commons talk:Structured data#Adding 'captured with' statement by bot Borealex (talk) 23:06, 24 May 2025 (UTC)

I've already been doing this for a while and some other bots too so we already have 40 million files with captured with (P4082).

Actually these edits will make it possible for the bots to pick up these files.

Your code looks like one of my old bots before structured data on Commons was added to Pywikibot. Have a look at a more recent one to see how you can do it now much cleaner. Multichill (talk) 18:43, 26 May 2025 (UTC)

Hi! My code is based on the code from here, and it's pretty old, yes. Thanks for the link, I'll study it soon. Borealex (talk) 08:43, 27 May 2025 (UTC)

I updated the code using pywikibot and the example from your bot. Thanks again for the help. Here is a test run after the update. Borealex (talk) 21:05, 2 June 2025 (UTC)

Ha, that looks much cleaner. You might also adopt the SPARQL logic so that you don't have to do that all manual. Multichill (talk) 17:00, 3 June 2025 (UTC)

Yes, I've already evaluated this script from your bot. At the moment, my bot's logic does not require a SPARQL-query, because as part of a each bot launch, the list of files will belong to a specific camera with a known QID. I plan to remove some of the manual work in the future by dealing with SQL queries, but I think I'll do it in the next step of updating my bot. Borealex (talk) 20:58, 4 June 2025 (UTC)

Rkieferbot (talk · contribs)

Operator: Rkieferbaum (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought:

• Add {{GPS EXIF}} to files that contain valid coordinates in EXIF but no location tags;
• Add Category:Ambiguous location in EXIF to files that contain invalid location in EXIF;
• Add Category:Taken with <Camera> to files that don't have it but contain valid and identifiable Make/Model properties in EXIF;
• Replace {{GPS EXIF}} with {{Location}} tags after at least a week of a file having it and after manual review of files in Category:Media with GPS EXIF.

Automatic or manually assisted: Semiautomatic (supervised).

Edit type (e.g. Continuous, daily, one time run): Weekly (after an initial drive to clear backlog.

Maximum edit rate (e.g. edits per minute): 12, usually fewer.

Bot flag requested: (Y/N): Y

Programming language(s): JavaScript for execution and python for building lists (until I'm comfortable and the logic is tested enough to run everything from python).

Rkieferbaum (talk) 16:49, 28 April 2025 (UTC)

Discussion

• Please make test run. --EugeneZelenko (talk) 15:40, 29 April 2025 (UTC)

  @EugeneZelenko: I already ran a few edits on the bot account, and the idea, at least initially, is to do pretty much what I've been doing in my main account, so feel free to take a look at my last few hundred contributions as well. Cheers. Rkieferbaum (talk) 16:35, 29 April 2025 (UTC)

• How many users are actually workung on manually checking {{GPS EXIF}}? How many cases have you seen where this was processed within a week? If nobody does it, the intermediate step would be just a waste of ressources? --Krd 07:09, 11 May 2025 (UTC)

  @Krd: I initially added location tags directly from EXIF, but this led to friction a few times. Either because other users thought a given photo belonged in Category:Location not applicable or because they knew the location information in EXIF was wrong. Users would sometimes remove the added location tag, but SchlurcherBot will sometimes have added that location to SDC, and not all users will correct that or remove it. So I changed the contents of {{GPS EXIF}} to invite watchers to indicate if one of those two situations applies, and that can be done before location is added. Since I switched to this two-step process, I've been following tagged photos and I'd say that on maybe 0.5% of them, watchers will either add a location tag, "Location not applicable" or "Ambiguous location in EXIF" tags. This is in addition to me (and whoever else wants to) manually checking Category:Media with GPS EXIF before running the script to add location to those. I feel like this is a good balance between prudence and efficiency, but I'm happy to make adjustments. Rkieferbaum (talk) 14:14, 13 May 2025 (UTC)

  How many are the mentioned "few times"? The whole request appears like a lot of manual work to me. Template:GPS EXIF currently has uncountable tranclusions, and Category:Media with GPS EXIF has 28k entries. I don't see that anybody really works on this. Either we can bulk place GPS EXIF by location template, or we cannot because there are too many false GPS data. Perhaps the last question should be answered in a village pump discussion. Am I mistaken? Krd 06:37, 29 May 2025 (UTC)

  @Krd: As I said, I'm open to skipping the intermediary part, but I do feel the extra (machine) work is small compared to the reduction in friction. Some context: I've zeroed the backlog in Category:Media with GPS EXIF several times, and pretty much all the 28k images currently sitting there were tagged by me over a few days. The only reason I haven't worked on that yet is that I'm hoping to do it via bot and therefore make it easier for others to manage their watchlists. I've spent about two months analyzing photographs with location in EXIF and no location tags, and the process I'm proposing to automate through this bot is essentially the same that I have been doing using my regular account. So I feel getting hung up on whether the intermediate step is necessary or not here is a little beside the point. It's a matter of whether my estimated 0.5% of edits following a GPS EXIF template addition is worth the extra step of tagging them before adding a location. My experience says it is. The extra work is done by a computer anyway. I'm happy to run that through the Village Pump, but, frankly, I don't quite see where the controversy lies. Rkieferbaum (talk) 00:51, 4 June 2025 (UTC)

  I suggest you make an extended trial with the bot account. Feel free to run at full scale is you think it makes sense. If nothing arises after some time, I tihnk this can be approved. (Bot flag set.) Krd 16:44, 12 June 2025 (UTC)

Железный капут (talk · contribs)

Operator: MBH (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions), Iluvatar (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions), Well very well (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought: Hello! As one of the co-maintainers of the bot I request for it bot and rollback flags. The bot detects suspicious potentially vandalous edits (on Commons this also includes copyright issues), streams them onto a Discord server, and trusted users revert/rollback/RfD them through an interface. Bot is already running on Russian, Ukrainian, Belarusian Wikipedias and Wikidata.

Automatic or manually assisted: Manually assisted

Edit type (e.g. Continuous, daily, one time run): Continuous

Maximum edit rate (e.g. edits per minute): Few edits per day

Bot flag requested: (Y/N): Y

Programming language(s): C# and Python

Well very well (talk) 06:16, 20 March 2025 (UTC)

Discussion

@CptViraj Seems like for some reason here on Commons, when a user is linked in bot edits' descriptions that user gets pinged -- a behaviour that doesn't happen on ru/uk/bewiki or WD... Do you know what it may be caused by and would it be fixed with the bot flag? Well very well (talk) 13:01, 23 March 2025 (UTC)

@Well very well: Per T189040, notifications aren't sent for edit summary mentions if the edit is marked as bot edit (b), so yeah, this should be fixed with the bot flag. -- CptViraj (talk) 16:00, 23 March 2025 (UTC)

It seems from subtasks of that task that just adding : to the start of link should remove the ping. I will go currently with this approach then. Well very well (talk) 17:32, 23 March 2025 (UTC)

Edits adding speedy deletion tags should generally not be marked as bot, as they should not be hidden from watchlists. AntiCompositeNumber (talk) 19:06, 24 March 2025 (UTC)

Currently bot doesn't mark its edits as bot at all, so this shouldn't be a concern. Well very well (talk) 15:34, 25 March 2025 (UTC)

I have some concerns. The test edits that have been made are mostly speedy deletion nominations, which is outside of the bot task you have described, especially since the majority of the files were tagged for copyright reasons, not vandalism. Please revise the request to accurately describe what the bot is designed for and capable of doing. Additionally, the bot does not notify page creators of the speedy deletion tags. More broadly, this bot task appears to mostly have the effect of removing rollbacks and speedy deletion tags from a user's contributions & deleted contributions, where they can be easily monitored by administrators. This is important both to deal with bad reviewing and to be able to establish a history of accurate copyright tagging when applying for advanced rights. And while rollback-like tools like SWViewer and Twinkle make the Commons rollback right less important than it might have been in the past, this bot gives administrators no control over who may be using it. AntiCompositeNumber (talk) 19:05, 24 March 2025 (UTC)

@AntiCompositeNumber
Please revise the request to accurately describe what the bot is designed for and capable of doing.
Done.
Additionally, the bot does not notify page creators of the speedy deletion tags.
Ok, will be implemented, thanks!
More broadly, this bot task appears to mostly have the effect of removing rollbacks and speedy deletion tags from a user's contributions & deleted contributions, where they can be easily monitored by administrators. This is important both to deal with bad reviewing and to be able to establish a history of accurate copyright tagging when applying for advanced rights.
Well, you can use tool for searching through user's descriptions — for that reason the bot always includes the type of action (RfD/rollback/undo) and user who did it in its edit description.
And while rollback-like tools like SWViewer and Twinkle make the Commons rollback right less important than it might have been in the past, this bot gives administrators no control over who may be using it.
Hm... @MBH @Iluvatar Maybe we can make the bot take the users list from a wiki page, e.g. a subpage of bot userpage on Meta? In this case everyone could see the list and stewards could also edit it. Well very well (talk) 15:42, 25 March 2025 (UTC)

@Well very well: Please report current state. --Krd 07:03, 11 May 2025 (UTC)

@Well very well:  ? --Krd 06:22, 29 May 2025 (UTC)

Oops, sorry, I wasn't checking the pings. I wasn't really working on the bot in the last 2 months (will continue the work soon) — so everything's the same currently. Should I proceed with implementing the suggestion from my last comment? Well very well (talk) 11:39, 29 May 2025 (UTC)

I'm not sure, but in any case please keep this page updated as long as the request is open. Krd 16:37, 12 June 2025 (UTC)