User talk:JarektBot/2020

original creation by uploader

Hi, this bot currently flags lots of my old files incorrectly as "original creation by uploader". Most of them were only uploaded by me, but not created. The author is different from the uploader. Suggest to restrict this task to files that use the {{Own work by original uploader}}. -- Schlurcher (talk) 12:13, 25 December 2019 (UTC)

Currently blocked until this is fixed. Thanks. Rodhullandemu (talk) 16:46, 25 December 2019 (UTC)

@Schlurcher and Rodhullandemu: I paused this quickstatements batch. Can you provide some examples of the issues? The database query I use (https://quarry.wmflabs.org/query/40759 ) specifically only looks for files that transclude {{Own}} template and this property is 100% equivalent. --JarektBot (talk) 17:10, 25 December 2019 (UTC)

I do disagree that "own" and "own work by original uploader" are the same. Please see for example File:Phantasialand_condor.jpg. This also applies to all files transfered from wikipedia. --Schlurcher (talk) 17:33, 25 December 2019 (UTC)

The wording on {{Own}} reflect only the majority of the use of the template. The template is also used for cases where the "own work" applies to the author not the uploader. As far as I see, JarektBot is also not approve for the performed task. Thus, my suggestion would be to get community input and discuss this properly as part of a new request on Commons:Bots/Requests. Part of this discussion should include, how the edits done so far can be corrected. --Schlurcher (talk) 19:22, 25 December 2019 (UTC)

I did check the 46'996 edits from your bot for the condition that uploader name is not included in author name. This gave me a list of 8'917 hits. I did check some manualy and in ~90% of the cases a different spelling or the full name was used in the author field, but based on reasonable guessing they are the same person. --Schlurcher (talk) 14:37, 6 January 2020 (UTC)

Schlurcher, I was thinking more about your point and although I still think that SDC property I was adding is equivalent to {{Own}} template, I realized that I should not rely on people always using that template properly. For example, funny fact is that I was the one who added it to your example file File:Phantasialand_condor.jpg in 2009. So in the future, if I add it, I will only add it to files where uploader and the listed author are the same, so that we will not propagate the problem. I will also check the files I already added it to. --Jarekt (talk) 03:35, 27 December 2019 (UTC)

I also did some further thinking about this (and performed a database search). Currently, like 48'000'000 files use the {{Own}}. I tend to agree with your point that the use of own has shifted. If I understand correctly, your Bot started from the genesis of Commons, so it would pick up the earlier (more uglier) uploads first. Your approach to check if uploader is equal to the listed author seems reasonable, but based on a quick check that I performed it will not cover all cases. There are also 1) people that give their full name in author but use a nickname for the account, 2) there are people that have changed their nickname and I think wikimedia has updated the uploader name but the name on the author field gets only forwarded. I do not want to slow down your vigor, but I think it is worth discussing this as a new bot request. Sometimes there is a better solution (or I might be wrong with the use of own and we could chatch these as people will complain as I did :-). A discussion at bot request will also provide details for these people. I was also thinking if there is a better approach than editing millions of files. I'm not really involved in the structured commons initiation, so maybe you can help. Basically you intend to add a property to all files that use {{Own}}. Can't we add the property to the template instead and it gets automatically added to each file (like inherited)? Is this possible? Because the next step is that every license file info will be added as a property and this would require the next bot action on each file on commons, again. I also think that in the long run we want also basic picture info (like size, resolution, date, camera, etc) as parameters. If there is a solution to inherit this from templates or file parameters this would make live easier. --Schlurcher (talk) 11:05, 27 December 2019 (UTC)

Schlurcher, My "bot" is not really a bot, I just use my bot account for large volume edits. All I was doing was to call SQL query quarry:40759 to generate QuickStatements commands which were then executed by QuickStatements tool. I had to work in batches of ~20k files so to do 48M would take a LONG time, but you have to start somewhere, and swarm of small simple "bots" is one way to do it. The SQL query with limit of 20k did seem to start with the older "uglier" files. My new proposed approach is not doable, based on my discussion on mw:Talk:Quarry since there is no way to get "author" based on the SQL query, so my approach for adding this property will need a proper python "bot". SDC properties can not be added by templates, and yes it will be a lot of edits to add properties for authors, licenses, sources, locations, etc. --JarektBot (talk) 14:19, 27 December 2019 (UTC)

@Jarekt: , hope you don't mind me bothering you again. So far I did not start looking into SDC, primarily due to lack of api/tool support. I'm basically waiting for a pywikibot implementation for wikidata/structured commons (with proper documentation). Your link to Quickstatements (which I was not aware of) made me want to look into this again. It looks like a nice tool, so I did some first testing. But basically, you need to provide a properly currated list for it to work on. I also noted that Quickstatements will add another statement value, even if the statement value already exists. Based on my prior comments, I think a list that gives all pages that use {{Own}} may include some incorrect cases. I'm, however, able to pull the uploader, author, and media ID information from the Commons API. This would allow me to perform the checks you suggested above as well as to check if Q66458942 is already availible.

I would like to start with re-checking the files that your bot already processed. Do you have a list of all files that you edited? Preferrably a list of all files that you tried to edit (including potential cases that did not lead to an edit). Otherwise, I think I can pull from the page edits from the bot. Please let me know. Thanks --Schlurcher (talk) 12:02, 3 January 2020 (UTC)

Schlurcher I am a big fan of QS and I used it extensively in wikidata. Unfortunately, although the tool worked for me for a while on Commons, now it does not seem to. Also the commands used for statement removal do not seem to work, see my Commons_talk:Structured_data#Problems_with_QuickStatements2_when_interacting_with_SDC post. It is quite frustrating as this tool is my main access to SDC, and it is unclear if there are any plans to fix it. In the mean time I did analyzed my edits adding own statements and I do have a list of couple thousand I would like to remove. My analysis consisted of a database query, like quarry:query/40838 to find original and final uploader of each image and a little AWB scraping script to capture "author" field of all the files. Comparing the two gave me a list of the cases where I want to remove my SDC statement. Unfortunately, I was unable to do the removal with QS tool. I am looking at some of user:Multichill python codes, like this one to see if I can get them installed and than adopted to do other tasks, like removal of statements. It has been many years since I run any pywikibot, but I do work much more with python those days, so I am hopeful. --Jarekt (talk) 14:30, 3 January 2020 (UTC)

Because we still have a ton of work to do, that prototype bot is quite conservative. It looks for both {{Own}} and {{Self}} (without author). Still you could run it this whole year and it would still have work left for next year. It's easier to make it more broad later than having to clean up. I'll it up again to at least get the easy cases converted. Multichill (talk) 18:00, 3 January 2020 (UTC)

FYI, I have added a request to Commons:Bots/Requests/SchlurcherBot8 to get some feedback from the community on the planned actions. --Schlurcher (talk) 14:26, 6 January 2020 (UTC)

This should stop AWB

Hi Jarekt, This seems wrong [1]. Please have a look. --Schlurcher (talk) 18:22, 27 January 2020 (UTC)

Schlurcher Wow, I was just replacing calls to Module:Date with equivalent calls to {{Date}}, using standard tool. That should not have happen. I will investigate. Thanks. --Jarekt (talk) 18:57, 27 January 2020 (UTC)

So strange, File:Beach volleyball at the 2018 Summer Youth Olympics – Girls Quarterfinals – ESP vs USA 1081.jpg was the last correct edit in a single run followed by File:Beach volleyball at the 2018 Summer Youth Olympics – Girls Quarterfinals – ESP vs USA 1082.jpg which blanked the page. The content of the page seems the same and my simple find an replace did not even have any regular expressions. I think that was some sort of Database error. Anyway I am fixing as we speak. --JarektBot (talk) 19:51, 27 January 2020 (UTC)

Schlurcher should be all fixed now. None of the files in Category:Files with no machine-readable license were recently edited by me. --JarektBot (talk) 23:03, 27 January 2020 (UTC)

Thanks for fixing this. It was a pure chance finding on my end. I reviewed the recent changes for something else and saw the edits from your bot marked with with the tag cleared the page which got my attention. --Schlurcher (talk) 09:32, 28 January 2020 (UTC)

QuickStatements

Hi, nothing too serious, but while I have set in my Preferences "Hide bot edits from the watchlist", I still have in my watchlist a substantial number of notifications regarding this kind of edit. I repeat it's not too serious, but I wanted to notify that to you. Regards, Christian Ferrer ^(talk) 17:47, 19 April 2020 (UTC)

Christian Ferrer, I do not fully understand it, but it seems like even if my bot has a bot flag, my bot's edits are not flagged as bot edits. I was searching phabricator for possible solutions and just found bunch of people reporting the same issue, which were closed as wont-fix. I think it is an issue with OAuth system or with QuickStatements. I was thinking about creating another ticket but it is unlikely to be fixed. --JarektBot (talk) 19:08, 19 April 2020 (UTC)

Ok, thanks Christian Ferrer ^(talk) 19:21, 19 April 2020 (UTC)

Список наблюдения

Русский: Добрый день. У меня в списке наблюдения очень много ваших правок. При этом в настройках отключён показ правок ботов. Правки других ботов скрываются, однако правки вашего бота каким-то образом попадают в мой список наблюдения. Прошу исправить это, чтобы правки вашего бота не показывались в моём списке наблюдения. Спасибо.

--Mitte27 (talk) 19:38, 20 April 2020 (UTC)

Mitte27, I am sorry but I have no control over which edits show up in people's watchlist. I am using standard tool d:Help:QuickStatements, which uses OAuth, see phabricator:T246746 for relevant discussion. --JarektBot (talk) 19:52, 20 April 2020 (UTC)

Русский: Разве нет никакого способа скрыть эти правки?

--Mitte27 (talk) 19:55, 20 April 2020 (UTC)

I am sorry but from what I read at phabricator, I do not see a way. --JarektBot (talk) 19:58, 20 April 2020 (UTC)

Malfunctioning

@Jarekt: I've blocked the bot because it was blanking a bunch of file pages with the edit summary "nil edit using AWB". Feel free to unblock once the error is corrected. Cheers, Pi.1415926535 (talk) 21:38, 24 April 2020 (UTC)

Pi.1415926535, thanks. With AWB bots leaving a message on the bot's talk page is enough to stop it. I hate when that happens. The bot was purging all the pages in Category:Pages with script errors for last few days, it got about 80k pages done and then I guess something went wrong. I will start fixing my edits. --JarektBot (talk) 22:38, 24 April 2020 (UTC)

Ok I believe I fixed all the issues. --Jarekt (talk) 23:19, 24 April 2020 (UTC)

Good to know - I'll remember that! The maintenance work you do with your bot is very much appreciated. Cheers, Pi.1415926535 (talk) 01:00, 27 April 2020 (UTC)

bot flag not set?

Hello, please check if your tool does set the bot flag correctly. As far as I see it is not set. Thx. --Krd 09:20, 4 May 2020 (UTC)

Krd The issue is discussed at Phabricator:T246746, but it is unclear if this is QuickStatements issue, "asynchronous" batch issue (whatever it is, it is used by QS) or OAuth issue. I would like to finish adding OTRS numbers this way, but I will not start on another property until this is fixed. --JarektBot (talk) 12:33, 4 May 2020 (UTC)

How many edits are left in the current batch? I personally don't mind as I have bot edits enabled on my watchlist anyway, but there may be users who disagree. --Krd 05:45, 5 May 2020 (UTC)

It is hard to say. I am trying to empty Category:Files with PermissionOTRS template but without P6305 SDC statement, which today has 47k files, however it takes a while for files to show up there. The total number is all the files in Category:Items with OTRS permission confirmed or 1.3M. So far I did about 20 batches of 25k so about 500k files so I have about 800k to go. --JarektBot (talk) 13:25, 5 May 2020 (UTC)

Category:San Francisco Bay Salt Ponds

There was an error in the upload. FOr the 124 files I uploaded today:

https://commons.wikimedia.org/w/index.php?title=Special:Log&offset=20200515000000&limit=124&type=upload&user=Evrik&page=&wpdate=2020-05-14&tagfilter=&wpfilters%5B0%5D=newusers&day=14&month=5&year=2020

Can the bot add this category Category:San Francisco Bay Salt Ponds? Thanks. Evrik (talk) 01:44, 15 May 2020 (UTC)

It looks like you figured out how to do it already. --JarektBot (talk) 03:22, 16 May 2020 (UTC)

Digital representation of, croppings, bot work

Hi. Your bot is constantly adding digital representation of to a cropping (not a faithful digitized representation of the indicated object or work, since it only covers 14.6% of the original artwork). Is there anything that can be done? Strakhov (talk) 13:20, 26 June 2020 (UTC)

Good point, the statement just means that content of the image is a digitization of an artwork in a given item. But it would be nice to separate full artworks from cropped artworks. I agree that we should add a qualifier to mark some of them as croppings. --Jarekt (talk) 18:34, 26 June 2020 (UTC)

Well, I thought (a least with regard to paintings) the property was intended to tag full pictures with professional quality (taken by museum employees => canonical representations). If that were true, maybe programming the bot to not add P6243 to files including "(cropped)" as part of their filename (or the {{Extracted from}} template) will sort out a few cases. If it's like you say and it's OK using it this way... I don't know what qualifier will fit. P2868? Strakhov (talk) 19:18, 26 June 2020 (UTC)

The way the property is used is that file with P6243 and {{Artwork}}, {{Art photo}} or {{Book}} template is treated identically to file with the same template with "Wikidata" field. So every file with {{Artwork}} where we have "Wikidata" field I add P6243. --Jarekt (talk) 02:33, 27 June 2020 (UTC)

To get an image license

I want to get a license from a photo. Will help please. I have no knowledge about OTRS. So please help Mr. Wiki Indian (talk) 03:04, 18 July 2020 (UTC)

Mr. Wiki Indian, If you are the photographer that took a photo than you can upload it to commons, but if you upload someone else's photograph than the photographer need to send permission for the publication to OTRS. I would advice to upload only photographs you have taken untill you get more familiar with Commons. --JarektBot (talk) 18:59, 20 July 2020 (UTC)

JarekBot, I have uploaded a photo (File:Amit Bhadana Instagram.jpg) on which I do not have rights. I ask for permission from OTRS for that. But I do not know how to request. That's why I'm telling you. help please. Mr. Wiki Indian (talk) 04:09, 21 July 2020 (UTC)

Mr. Wiki Indian, There is not much you can do. If File:Amit Bhadana Instagram.jpg is not your photo than it will be deleted. In rare cases you can contact the photographer who took that photo and he (or she) can send permission to use it to OTRS. But in most cases it is unclear who took the photo or they would not agree to to release it. --Jarekt (talk) 04:16, 21 July 2020 (UTC)

Cleaning up

Is this cleaning up really usefull? Much information is destroyed. --Wikiwal (talk) 17:20, 25 July 2020 (UTC)

Adding to this that this also happened on several other files that I'm currently cleaning up, see my contributions. Please check why this happened, it's clearly not intended. Tokfo (talk) 19:33, 25 July 2020 (UTC)

Wikiwal and Tokfo Thank you for quickly notifying me. Leaving a message at this talkpage stops AutoWikiBrowser (AWB) "bot" task. I use AWB a lot (5M edits so far) but once in a blue moon AWB goes berserk and starts blanking pages. I reported this issue to the authors, but it is so rare that they never managed to reproduce it yet. I think I undid all the blanking by now. Thanks --Jarekt (talk) 02:27, 26 July 2020 (UTC)

Very good! Thanx a lot, --Wikiwal (talk) 06:34, 26 July 2020 (UTC)

Bot flag

Mass automated changes like diff are not being marked as by a bot. Please correctly mark these with the bot flag to avoid filling up RC/watchlists. Thanks --Fæ (talk) 07:14, 26 July 2020 (UTC)

It is a known issue, likely with QuickStatements tool, discussed in phabricator:T247433 and phabricator:T246746. Please add your support for those tickets and maybe someone will look into it. According to User:RIsler (WMF) (in one of those tickets) this is on "back burner for now and see if it becomes an issue". However I am also not sure what WMF can do as the issue is likely with the QuickStatements tool. --Jarekt (talk) 13:27, 27 July 2020 (UTC)

Hello. I have the same issue in my watchlist and I do not know how to "support a ticket". Do you think it could be possible to stop your bot until the issue is fixed? Thanks. --Le Petit Chat (talk) 21:15, 28 July 2020 (UTC)

By supporting the ticket I mean, taking part of the discussion there so people prioritizing tickets know that this is important. I am sorry about this annoyance and I do not expect this issue to be solved anytime soon. My watch list is also often flooded by my own bot and there is not much I can do about it. --JarektBot (talk) 00:53, 29 July 2020 (UTC)