User talk:Ivan A. Krestinin

From Wikidata
Jump to: navigation, search
On this page, old discussions are archived. An overview of all archives can be found at this page's archive index. The current archive is located at User talk:Ivan A. Krestinin/Archive.

Contents

Constraints[edit]

That's great we are migrating to property statements! I had already started a module but then let it be because I didn't know how it should look like. Is there anything you would use in your system? I'm asking just because it is obsolete now and I will delete it.

Do you also plan migrating using bot? Matěj Suchánek (talk) 14:28, 30 November 2015 (UTC)

I have no enough LUA experience to understand all features of your module. Migration tool/bot is ready. But one issue blocks this process now. — Ivan A. Krestinin (talk) 15:00, 30 November 2015 (UTC)

НП-Беларусь[edit]

Добрый день! Можете ботом добавить на Викиданные все статьи из этой категории? --Чаховіч Уладзіслаў (talk) 12:29, 1 December 2015 (UTC)

  • Приветствую, по категории мне не очень удобно. Запустил бота по всем статьям из be_x_oldwiki, bewiki и ruwiki, которые содержат no label (Q5625611). Надеюсь это покроет категорию. — Ivan A. Krestinin (talk) 20:26, 1 December 2015 (UTC)
Да, можно и так. Раз уж Вы вытягивает всё из шаблонов в статьях, то может и названия из шаблона lang-be-trans вытяните в качестве дополнительных для крупных языковых разделов, которые используют латинский алфавит (en, de, fr, it, es, sv, pl, pt, можно еще lt, lv, et)? Это бы упростило поиск и объединение статьей с категориями на Викискладе, например. --Чаховіч Уладзіслаў (talk) 21:19, 1 December 2015 (UTC)
Не уверен, что это правильно, всё-таки названия на разных языках нередко пишутся по-разному. — Ivan A. Krestinin (talk) 16:30, 4 December 2015 (UTC)
Поэтому я и не предлагаю добавлять эти названия в качестве основных. К сожалению, свойства "транслитерация" в Викиданных нет (либо я не знаю про него).--Чаховіч Уладзіслаў (talk) 21:53, 4 December 2015 (UTC)
Неосновные названия тоже должны быть написаны на определённом языке. Не хочется мне что-то такое делать, по-моему это не слишком правильно. — Ivan A. Krestinin (talk) 22:10, 4 December 2015 (UTC)
Создание элементов кстати бот закончил, плохо что в статьях нет какого-нибудь уникального идентификатора, какого-нибудь аналога кода ОКАТО, ОКТМО или хотя бы почтового индекса, а то можно было бы автоматически объединить элементы. — Ivan A. Krestinin (talk) 19:44, 5 December 2015 (UTC)
Благодарю! Есть список всех-всех населенных пунктов Беларуси (в том числе ликвидированных) с указанием кода Code System for Administrative-territory Division Objects and Settlements (Q6538410), но, к сожалению, он содержит только русские названия и немного устаревшее административное деление, поэтому пока что затруднительно использовать информацию из него для заливки в статьи в bewiki, хотя один участник взялся его адаптировать. Также на сайте Госкартгеоцентра есть странички для каждой административной единицы (и не только) Беларуси (пример - Вауки, искать можно здесь). Пока могу выставить заявку на создание соответствующих свойств. --Чаховіч Уладзіслаў (talk) 10:38, 12 December 2015 (UTC)
Свойства — это хорошо, только если в статьях не указаны эти идентификаторы, то проставить их ботом будет довольно сложно. — Ivan A. Krestinin (talk) 17:53, 12 December 2015 (UTC)

Expansion depth limit exceeded Constraint:Units[edit]

Your latest edit to {{Constraint:Units}} made it impossible to use {{Q}} for the values for list. It now gives a Expansion depth limit exceeded. See for example http://www.wikidata.org/w/index.php?title=Property_talk:P2370&oldid=278251799. Could you check how to fix this? This makes pretty display of units impossible. Mbch331 (talk) 15:56, 1 December 2015 (UTC)

It is strange behavior, I have no idea how to fix this:
  • {{#invoke:Constraints|getCaption|units|123456}} has expansion depth 2.
  • {{Q|123456}} has expansion depth 6.
but
  • {{#invoke:Constraints|getCaption|units|{{Q|123456}}}} has expansion depth 41.
I do not understand why 2 + 6 = 41. If you do not have any ideas too then please rollback my edit. The edit is not important. — Ivan A. Krestinin (talk) 20:48, 1 December 2015 (UTC)
I have no idea either how it's possible. So I reverted your last edit. But it might be worth to let someone with more knowledge of lua take a look. Mbch331 (talk) 21:16, 1 December 2015 (UTC)

Database reports[edit]

Hi Ivan, just found your Wikidata:Database reports/Identical VIAF ID. This list is a great help. Is there a simular list for GND ID (P227)? If not, can you create one? --Kolja21 (talk) 01:29, 3 December 2015 (UTC)

Awesome. Thanks! --Kolja21 (talk) 17:05, 7 December 2015 (UTC)

When Mohammed do not go to the mountain![edit]

I would like to invite you to sv:WP:WF#Wikidata to responde to questions about KrBots edits to the VIAF-property! I invited Annika64 to come here to talk, but she says she do not want be in such discussions! Fell free to write in English! -- Innocent bystander (talk) 19:31, 3 December 2015 (UTC)

Revert[edit]

FYI Q319537. Please double check. --- Jura 12:58, 8 December 2015 (UTC)

Constraint Type[edit]

Здравствуйте. Скажите пожалуйста, почему "Constraint:Type" здесь не работает? Danneks (talk) 12:24, 12 December 2015 (UTC)

Constraint is defined to check the value for subclass of (P279) on the item that uses statement describes (P2384). However in the violation cases there is no subclass of (P279). So either the constraint needs to be changes or the items that violate the constraint. Mbch331 (talk) 12:42, 12 December 2015 (UTC)
Thank you. I'm not quite sure which solution is right, but it is easier to change the constraint. Danneks (talk) 12:55, 12 December 2015 (UTC)

Vandalism[edit]

Your bot seems to add data without any human verification. Please stop this vandalism... Sumurai8 (talk) 16:23, 12 December 2015 (UTC)

@Sumurai8: The source is stated for the edits. Wikidata works a bit different than Wikipedia. Importing data from Wikipedia's is currently still accepted. As long as the community doesn't say it's not allowed, it's still allowed to import data from Wikipedia. Mbch331 (talk) 16:33, 12 December 2015 (UTC)
Thank you, Mbch331. I can skip dates from future. But most probably this case is single. Vandalism detection problem does not have robust general solution unfortunately. — Ivan A. Krestinin (talk) 17:51, 12 December 2015 (UTC)
I think it's a single issue too, but to be safe it is indeed best to skip future dates for birth and death. Mbch331 (talk) 18:05, 12 December 2015 (UTC)
It would help if you used the latest tested/checked version of wikipedia's that uses that. You imported an unchecked vandalised page edited by an anonymous user. The tested/checked version was correct. In the worst case the tested/checked version is incomplete, but at least a human has read through it. Alternatively, you can possibly log pages where the current version is not the tested/checked version and manually check the content before parsing it and importing the data. Sumurai8 (talk) 18:36, 12 December 2015 (UTC)
Wikipedia:Reviewing (Q11107170)-based approach is looked reasonable and implementable. But unfortunately the most Wikipedias have too bad patrolling rate. The only exception is dewiki as I know. Additional think: popular well-patrolled articles are imported already. My bot is working with low popular articles now. Patrolling rate has additional decrease in this area. — Ivan A. Krestinin (talk) 19:34, 12 December 2015 (UTC)
These kind of errors show up in some reports, not the baddest vandalism we have here. Sjoerd de Bruin (talk) 22:12, 12 December 2015 (UTC)

Self-reference[edit]

Have you any idea why this edit was done by KrBot? -- Innocent bystander (talk) 17:48, 13 December 2015 (UTC)

Ok, thanks. That merge was probably wrong, but I am not in the mood to argue about it today. -- Innocent bystander (talk) 08:29, 14 December 2015 (UTC)

Error on marking church as human Q21706793[edit]

Hello. Please fix your bot, because of follow issue: https://www.wikidata.org/w/index.php?title=Q21706793&type=revision&diff=282684636&oldid=282684625 --Movses (talk) 09:39, 25 December 2015 (UTC)

Wikidata:Database reports/Constraint violations/P373[edit]

It is possible to fix that error? --Jklamo (talk) 22:19, 25 December 2015 (UTC)

  • Fixed, please wait the next reports update. — Ivan A. Krestinin (talk) 09:22, 26 December 2015 (UTC)
    • Thanks! --Jklamo (talk) 12:22, 26 December 2015 (UTC)

Возера и Рака[edit]

Здравствуйте, можете "выпотрошить" шаблоны Возера и Рака? --Чаховіч Уладзіслаў (talk) 08:34, 26 December 2015 (UTC)

Qualifier or reference constraint[edit]

Currently, there are constraints for value-only, qualifier-only, reference-only constraints. For P1683, a "qualifier or reference"-constraint might be useful. --- Jura 14:06, 28 December 2015 (UTC)

Adding information[edit]

Hi again, I wanted to share with you an idea how to make database reports even more useful. You know, in Types Statistics, the bot lists how many instances are there of each type. You could program the output with a link to AutoList2 which would then show the reader a list of all those instances. It's quite simple, AutoList2 has a permalink option, so it's only a matter of formatting the link correctly.

Like this: http://tools.wmflabs.org/autolist/index.php?language=en&project=wikipedia&category=&depth=0&wdq=claim%5B{{{1}}}%5D%20AND%20claim%5B31%3A{{{2}}}%5D&pagepile=&statementlist=&run=Run&mode_manual=not&mode_cat=not&mode_wdq=and&mode_find=not&chunk_size=10000, where {{{1}}} is the property reported and {{{2}}} is type.

For example, the Types Statistic in report for P1587 would look like this:

What do you think? — Yerpo Eh? 09:08, 29 December 2015 (UTC)

Without "language=en" of course. And may be "chunk-size=1000" as it loads quicker. --Infovarius (talk) 21:30, 30 December 2015 (UTC)
Many reports are very large. Adding many long URLs will increase reports size. Is the links really useful? What tasks require the links? — Ivan A. Krestinin (talk) 19:33, 31 December 2015 (UTC)
I think it's useful to see which items are of some type, but it's your decision. For the example I gave above, it's useful to cross-check with official list of monuments. — Yerpo Eh? 19:05, 2 January 2016 (UTC)

ISNI[edit]

The use of spaces in the ISNI number is disupted.. Please stop your bot.  – The preceding unsigned comment was added by GerardM (talk • contribs).

Constraint report update/KrBot[edit]

Hi Ivan,

I guess the bot has a day off .. dumps would be available.
--- Jura 11:37, 15 January 2016 (UTC)

  • Hello Jura, I had connection troubles. Bot will update all reports in the next 3-5 hours. — Ivan A. Krestinin (talk) 19:33, 15 January 2016 (UTC)
    • No hurry. Thanks Ivan.
      --- Jura 21:56, 15 January 2016 (UTC)

List/category of failed constraint violation reports[edit]

Hi! I'm new to Wikidata and I just happened to discover several property constraint violation reports that had been failing for some time now, e.g. [1] and [2] (I fixed all the ones I found).

That's all fine, people make mistakes. But I couldn't find any existing pages about failing reports. To aid fixing these problems, could you create a list page with links to all failed property reports? That way interested people can watch it and fix problems quickly when they show up.

If that's too complicated, maybe just include a template or add a category to failed report pages. The downside is that categories/transclusions can't be watched using the watchlist (AFAIK?), but it's still better than having to search for failed reports. Intgr (talk) 11:50, 28 January 2016 (UTC)

Constraint violations/P1447[edit]

Hi! It looks like I screwed up it a little bit :) Maybe your bot could recognise {{Q|Qx}} as a valid expression because {{Q}} accepts it? --Edgars2007 (talk) 11:13, 29 January 2016 (UTC)

  • Hello, it can be supported, but it is looked redundant. Next step on this way is looked as {{Q|QQx}} or {{Q|Px}} :). I created new report: Wikidata:Database reports/Constraint violations/Errors for catching such errors and another issues. — Ivan A. Krestinin (talk) 18:44, 31 January 2016 (UTC)
    • Well, the main difference between {{Q|Qx}} and {{Q|QQx}} (and {{Q|Px}}) is that the first one works, but other two doesn't :) But of course, it's up to you, it was just a suggestion. --Edgars2007 (talk) 19:45, 31 January 2016 (UTC)

Дубликаты[edit]

Здравствуйте. Нашёл в русской википедии ~700 статей про украинских сельсоветов, которые имеют свой личный элемент Викиданных и не связаны с аналогичным статей из украинской википедии (а они существуют и тоже имеют свой элемент. т.е. итемы-дубликаты). Список здесь. Можете найти какой-нибудь удобный алгоритм для вашего бота что-бы найти и объеденить пары этих элементов? --XXN, 15:59, 9 February 2016 (UTC)

В похожем ситуации (но чуть сложнее) около 700 статей про украинских сёл в румынской википедии,[3][4] статьи которые оказались изолированы здесь в викиданных и нуждаются в связании с украинсками аналогами.
И если уже речь о дубликатов, заметив эту вашу инициативу, пользуясь случаем принесы к вашему вниманию факт что, на этот момент здесь в викиданные есть ~109.000 изолированных (одиноких) категорий немецкой википедии[5] и ~68.000 подобнух категорий французкой википедии[6]. Среди них конечно куча дубликатов для объединения, но проблема как бы найти им пару. Может сумеете зделать вы что нибудь с частью из них, а вообше-то при таком масштабе уже и немецкие девелоперы должны помочь, вовлечся как нибудь. Простите если побеспокоил. --XXN, 19:57, 9 February 2016 (UTC)
По сельсоветам попробую объединить используя КОАТУУ, практически готовый код для этого имеется. К концу предстоящих выходных наверняка сделаю, может раньше. С категориями надо ещё посмотреть. — Ivan A. Krestinin (talk) 04:53, 10 February 2016 (UTC)
@XXN: На данный момент сделал следующее: из шаблонов uk:Шаблон:Селищна рада, uk:Шаблон:Сільрада України, ru:Шаблон:Административная единица Украины импортировал KOATUU identifier (P1077), located in the administrative territorial entity (P131), украинское название. После этого объединил те элементы, у которых совпадали KOATUU identifier (P1077) и названия на украинском. Дополнительно проверял непротиворечивость located in the administrative territorial entity (P131). Завтра ещё раз пройдусь алгоритмом объединения по всем элементам, возможно что-то в сегодняшний дамп не попало. Румынской википедией пока занимаюсь, у них нет специализированных карточек для Украины, потому извлекать данные намного сложнее. Категориями не занимался. — Ivan A. Krestinin (talk) 12:06, 14 February 2016 (UTC)
По rowiki я смотрю ваш бот начал работать, скажите тогда, как он закончит, я тогда своим ботом ещё пройдусь, может ещё что-нибудь извлеку. Кстати у нас в метки похоже попали ударения, причём похоже не в том формате, в no label (Q4081102) метка "Безім'янка", в тоже время в rowiki указано название "Безім`янка". Не знаете случаем, это точно просто ударение, которое можно удалить из метки? Ещё у многих сёл в качестве located in the administrative territorial entity (P131) указан район, хотя в rowiki имеется и сельсовет, например no label (Q3617635), нужно такие случаи исправлять? — Ivan A. Krestinin (talk) 14:16, 14 February 2016 (UTC)
Про located in the administrative territorial entity (P131) - я не знаю точно какой здесь консенсус; наверное надо указывать низшее административное единица. Но посколько большентсво википедий не имеют статьи про сельсоветов, у меня лично на данный момент нету единой мнений что надо "исправить" ту значению из района в сельсовета, но при этом и не могу и не буду возражать если будем произведена однородность.
Да, как заметили, в последние дни я начал при помоши бота (а местами и вручную) объеденить изолированные итемы статей rowiki - а зделал это только там где мог автоматически/массово найти 100%-ые пары для румынских статей. И там было немножко хаоса: кроме тех более 1500 статей (если не ошибаюсь) про украинских Н.П. и сельсоветов имея собственный элемент WD, несвязанные с украинскими/русскими статьями-аналогами; были также, румынские статьи про сёл одиночки-в ихнем сельсовете - связаны неправильно с украинскими статьями про сельсоветов а не про село (это исправил до конца). Мой метод работы с ботом для слияний элементов не очень продвинут - нужно иметь точные паралельные списки элементов/названий статей; так что где я мог - формировал эти списки при помоши интрументов как PagePile, Tabernacle и Linked items, а потом приступал к слияний. Короче говоря мучался немало и на этот момент я исчерпал свой списки над которыми было поработать а остались ещё 1100 несвязанных статей - итемов для объединения [7]. Но думаю что вам может быть трудно незная румыснкий язык. В крацех, в rowiki статьи созданы по данным 2001 переписи и есть статьи и про исчезнувших сёл, для которых украинци не имеют аналог :), например про ru:Заречье (Белогорский район, Крым)ro:Zaricicea, Bilohirsk. Некоторые сёла, сельсоветы и даже один район созданы с старыми названиями (2001) а в украинском разделе (и остальные) у них уже новые, изменёные. А также, не знаю почему, не во всех статей rowiki указан код КОАТУУ, то что может вам создавать трудности. Мда, и похоже что мой коллега румынский ботовод как-то перепутал и поменял украинский апостроф с каким-то другим символом. Если вас ещё не испугали все эти трудности можете взяться и за этот таск, попробовать и посмотреть что получется. --XXN, 15:30, 14 February 2016 (UTC)
Накосячил этот ботовод не только с апострофом, но и с самими названиями. Большая загадка почему в качестве украинского названия в статье ro:Comuna Novomariivka, Solone значится Новомар`ївка вместо Новомар'ївська сільська рада. Похоже из rowiki украинские названия лучше и не пытаться импортировать. — Ivan A. Krestinin (talk) 20:25, 14 February 2016 (UTC)
Да, да. Как нибудь подскажу ему об этом. Дело в том что в румынском разделе нередко недооцениваются префиксы/дополнительные слова из словосочетаний названий админ. единиц. Вот например как выглядит верхушка статьи об Ясском уезде — "подзабылось" слово "Județul" и в вступлении и в карточке. Точно также и в случае коммун. --XXN, 22:21, 14 February 2016 (UTC)

Importing content from Wikipedia GeneWiki info boxes to Wikidata[edit]

Hi! Your bot seems to occasionally do imports from the English Wikipedia Gene Wiki infoboxes. On on hand, your bot seems to be malfunctioning, as it put the gene ontology term protein serine/threonine kinase activity (Q14326094) as molecular function (P680) and as biological process (P682) on item [8] and many other items (and also many other gene ontology terms), which is wrong. As we, the ProteinBoxBot team, have been importing large amounts of biological data from their primary sources, e.g. Uniprot and many of the data in the GeneWiki infoboxes are outdated, it might be the right time to stop importing data from the Wikipedia GeneWiki infoboxes. Furthermore, we are currently rolling out a new Wikipedia GeneWiki infobox template script, which pulls the data directly from Wikidata to populate these infoboxes [9]. Re-importing these data to Wikidata is therefore not required anymore and only creates a lot of overhead. More info on what we currently do and plan can be found here: [10]. Thank you! Sebotic (talk) 09:32, 10 February 2016 (UTC)

Oh, I just realized that your bot only fixed problematic redirects and is not responsible for the wrong Gene Ontology terms. Sorry for that! But in general, imports from GeneWiki infobox to Wikidata should not be required anymore. Best Sebotic (talk) 10:45, 10 February 2016 (UTC)

Undo[edit]

Hi Ivan,

Would you undo these changes. I fixed the corresponding merge.
--- Jura 15:52, 14 February 2016 (UTC)

  • Hello, Jura, bot is in progress now. I think we need to integrate different from (P1889) with MediaWiki:Gadget-Merge.js and another tools. Or we need to create another property pair: "do not merge with" and "reason". — Ivan A. Krestinin (talk) 18:32, 14 February 2016 (UTC)
    • When items you want to merge reference each other (one way is enough), it shouldn't be possible to merge them. So that this merge was possible is rather strange. Mbch331 (talk) 18:55, 14 February 2016 (UTC)

If you have an easy way to undo those merge-related edits, would you also undo those that followed the erroneous merger between Istrup (Q1674886) and defensive wall (Q57346) 8 Feb ? --Zolo (talk) 11:35, 27 February 2016 (UTC)

Confusing reference[edit]

Could you please explain what this edit is about? --Yair rand (talk) 12:05, 22 February 2016 (UTC)

Hello, bot detected duplicate values. After this bot copied all references to the first value and delete all another duplicate values. Please see the next bot`s edit. — Ivan A. Krestinin (talk) 12:32, 22 February 2016 (UTC)
Ah, I see. Thank you. --Yair rand (talk) 15:53, 22 February 2016 (UTC)


Year 0[edit]

Hi Ivan,

I undid a couple of these. They showed up with the "calendar model" complex constraint on P570.
--- Jura 13:10, 24 February 2016 (UTC)

Thank you[edit]

My thanks to you and your bot for [11] and [12] changes.

I put these with Magnus' Quick Statements, with a statement like:
LAST P569 +00000001946-08-03T00:00:00Z/11 P248 Q21087031
which is not putting the reference in the correct place.

Do you have any idea what would be the correct syntax to put it in the correct place? --FocalPoint (talk) 14:38, 20 March 2016 (UTC)

Do NOT change redirects[edit]

Your bot just made a wrong merge worse - instead of reverting just two edits I now have to find all those edits of your bot where it changed the link to the redirect to the new page, e.g. Bangkok Noi (Q806486). There is absolutely no technical reason to change any uses of redirected item. Ahoerstemeier (talk) 22:12, 1 April 2016 (UTC)

@Ahoerstemeier: quick answer, Ivan will give you more detailed answer. a) AFAIK, bot is able to revert these changes, you simply have to say which item has to be "reverted" b) this is not Wikipedia, there are reasons, why redirects should get fixed. There are 24h delay before doing that, so wrong merges can be undone. --Edgars2007 (talk) 22:23, 1 April 2016 (UTC)
Thank you, all right. Links are needed to be checked anyway after merge revert. Some links can be created during merged item existing time. — Ivan A. Krestinin (talk) 22:29, 1 April 2016 (UTC)
@Edgars2007: Help:Redirects#Links to redirects clearly states these links to redirect shall not be changed, which was also what I think to have seen as a discussion result in the project chat. Besides, 24h is way to short to spot a wrong merge if its not a prominent item. Ahoerstemeier (talk) 22:37, 1 April 2016 (UTC)
Help page was written without any discussion and situation analysis. — Ivan A. Krestinin (talk) 22:59, 1 April 2016 (UTC)


Merged items[edit]

Hi Ivan,

Just to let you know: I merged a series of items you created for articles a Kazakh wiki. They appeared on Wikidata:Database reports/identical birth and death dates/1.

I think it might be worth looking into Wikidata:Bot_requests#Cyrillic_merges. Obviously, it's not my area of expertise. Even if writing of names differs, I think some items could be identified by accounting for different name orders and punctuation. I keep coming across samples once they get dates.
--- Jura 16:56, 4 April 2016 (UTC)

Template:Constraint:Units[edit]

Hi again :)

area (P2046) is a generic property, so any unit of area should be allowed (we have 58 different units at the moment although only 10 are listed on the property talk page).

Would it be possible to have something like Template:Constraint:Value type? e.g. {{Constraint:Units|class=Q1371562|relation=instance}} to mean "the unit must be an instance of unit of area (Q1371562) (or one of its subclasses).

- Nikki (talk) 09:45, 10 April 2016 (UTC)

Hello, it is possible of cause. But are we really need to spend efforts to implement and support this ability? Only 10-11 units are used now: Wikidata:Database_reports/Constraint_violations/P2046#Units_statistics. The list can be extended at any moment. Also suggested approach is less stable. — Ivan A. Krestinin (talk) 10:42, 10 April 2016 (UTC)
It's not specific to this property, there are quite a lot of properties like this. In fact, I think most properties using units are more accurately described as having a type of unit (e.g. unit of length) as the allowed values for the unit. More examples:
It would be particularly useful for currencies - someone else has asked on Property_talk:P2284 if it would be possibly to just say any currency instead of trying to list all currencies.
- Nikki (talk) 10:44, 13 April 2016 (UTC)
Hello Ivan, I came here, because I had basically the same request. I just created budget (P2769) and it would be useful to say "unit can be any instance of budget (P2769)", instead of listing currencies by hand. --Srittau (talk) 12:44, 23 April 2016 (UTC)
In the meantime, we have 270 units of type "area". I don't think these are necessarily all suitable.
--- Jura 13:56, 23 April 2016 (UTC)

TLDR[edit]

I am slowly working with the P1566-report. One problem, that causes me to edit less in a day than I could, is that the report is lllllloooooonnnnngggg (857316 bytes in the last version). It takes me half a minute every time I successfully load the page, and far from every time is successful! The report is already shortened, not everything is written in the report. Can you make it even shorter? -- Innocent bystander (talk) 07:14, 20 April 2016 (UTC)

If you just want a small selection of the many violations to fix to work on, you would probably be better off using a SPARQL query (e.g. 50 items with P1566 and no country). You can even select labels or other properties that way. The SPARQL endpoint is also more up to date, whereas the constraint reports seem to be a couple of days behind (at least I often have to wait a couple of days before things I've fixed get removed from a constraint report and I can continue working on it). To be honest though, there's so much to fix for P1566 that fixing things by hand is really inefficient. The problem was caused by bots (mostly lsjbot importing GeoNames into svwiki and cebwiki as far as I can tell), so it should be possible to add lots of the missing statements with bots too. - Nikki (talk) 11:22, 20 April 2016 (UTC)
The errors added by Lsjbot are the most easy to solve, since every Lsjbot-article matches one and only one GeoNames-ID, while the man made articles have several topics and are difficult to match in our items. Many of the items in the unique-part of the list also have a wrong GeoNames-ID introduced by a local bot here. They have been imported directly from a GeoNames-dump. But the links to Wikipedia in GeoNames are often wrong, only matching the name of the article, not the topic.
I cannot see any other way than fixing this by hand. Adding P625/P17 can probably be added by bot, but such bot-runs have already been done, leaving a lot of items behind.
I am not familiar with SPARQL at all. Can you provide me a link for the unique-constraint? -- Innocent bystander (talk) 14:13, 20 April 2016 (UTC)
This should be what you need. Of course, we can add there (almost) anything else, that helps find violations better. P.S. Here is some help about SPARQL. --Edgars2007 (talk) 17:13, 20 April 2016 (UTC)
@Edgars2007, Nikki: The links to the WD-items in the SPARQL-report goes to json-versions of the pages, not to the same page links like Q1 takes you! -- Innocent bystander (talk) 06:28, 21 April 2016 (UTC)
That's the same problem you mentioned on Wikidata:Contact the development_team#Identify_the_unit. I've added a comment there. - Nikki (talk) 06:54, 21 April 2016 (UTC)
{{Q}} is overused on the report. The template will be excluded from long tables during next run. — Ivan A. Krestinin (talk) 18:49, 20 April 2016 (UTC)

ru:ПРО:ВО: бардак с реками[edit]

Приветствую!

Не поможешь навести порядок с реками?

  1. Для 53-х тысяч речных объектов в ВД не указано свойство страна: [13]
    1. Проблема, по-видимому, создана криворукими участниками, заливавшими от шведов (46 тысяч рек без стран): [14]
    2. Залить 800 штук можно из немецкого раздела: [15]
    3. Залить 1200 штук можно из английской вики: [16]
  2. В шведском разделе тысячи статей о реках с «ручными» интервиками. Элементы на ВД я им создал, а вот как туда привязать статьи из интервик, не знаю (там привязки к ВД ещё нет). Вот собственно и нужно привязать все статьи по «ручным» интервикам к объектам на ВД. Мастер теней (master of shadows), 18:27, 2 May 2016 (UTC)
  • У меня был какой-то алгоритм импортирующий страны, запустил его на шведском материале. Потом можно будет и по остальным пройтись. Чего-то готового для удаления интервик у меня нет, но у других ботоводов должно остаться со времён миграции на Викиданные. — Ivan A. Krestinin (talk) 15:06, 3 May 2016 (UTC)
Там можно и не удалять, в принципе. Главное связать с уже имеющимися объектами в ВД. Мастер теней (master of shadows), 16:04, 3 May 2016 (UTC)
У меня даже такого кода нет, статьи с локальными ивиками я обычно обходил стороной т. к. там велика вероятность неразрешённых ивики-конфликтов. Попробуйте спросить у участника Emaus, он в своё время занимался переносом интервик, если мне не изменяет память. — Ivan A. Krestinin (talk) 16:44, 3 May 2016 (UTC)
  • Бот прошёлся по странам, но получилось у него не везде. Бот не смог разобраться там, где в одном параметре была указана и страна и регион. Также он не справился там, где указано сразу несколько стран. — Ivan A. Krestinin (talk) 04:54, 6 May 2016 (UTC)

Ваш бот объединяет элементы, которые не должны быть объединены[edit]

Q17682271 и Q21672401 не должны быть объединены, это разные населенные пункты. --Alex NB IT (talk) 04:05, 4 May 2016 (UTC)

Вероятно потому, что кто-то указал одинаковые почтовые коды. Мастер теней (master of shadows), 08:54, 4 May 2016 (UTC)
Действительно, т. к. названия и почтовые индексы были одинаковыми, и противоречий в каких-либо других свойствах не было, то бот "решил", что это один и тот же населённый пункт. — Ivan A. Krestinin (talk) 09:23, 4 May 2016 (UTC)

Проставление кода ОКТМО[edit]

Ваш бот проставил код ОКТМО на страницу Q18769731. Похоже, эта страница - ошибка ботозаливки чеченской вики, никаких следов существования Васюково в Андреевском сельсовете или сельском поселении я не нашла. Единственное Васюково Вашкинского района это Q4105177 (и в нем правильно указан тот же код ОКТМО). Я бы стерла это значение ОКТМО из Q18769731, но не проставит ли его бот заново? --Daryona (talk) 10:18, 4 May 2016 (UTC)

  • Добрый день, похоже этот НП перестал существовать. Хорошо бы заполнить свойство "дата прекращения существования", тогда бот точно его трогать больше не будет. Но вот чем его заполнить - непонятно, я сходу тоже не нашёл упоминаний. Вообще с сёлами-призраками беда, их слишком легко перепутать с одноимёнными ныне существующими сёлами в этом же районе. — Ivan A. Krestinin (talk) 11:07, 4 May 2016 (UTC)
    • Да, я тоже сейчас нашла следы, что был такой НП раньше (в справочниках почтовых индексов, в частности, встречается). Но я выверяла по Вологодской области все НП с 1999 года, так что его еще раньше ликвидировали, и вряд ли удастся найти что-то сейчас. Может, указать "1999", чтобы просто была какая-то отметка об упразднении? Все равно вряд ли кто-то будет реально редактировать эту статью и страницу викиданных. --Daryona (talk) 12:49, 4 May 2016 (UTC)
  • День добрый. Бот начал проставлять коды ОКТМО и почтовые индексы в реки. Проблема с чеченской викой. Мастер теней (master of shadows), 14:11, 22 May 2016 (UTC)
    • Да, видел немало таких ситуаций, но тут уж ничего не поделать, нужно разгребать неправильные интервики. — Ivan A. Krestinin (talk) 14:29, 22 May 2016 (UTC)
Почему? Можно спрашивать P31. Если не подкласс НП — зачищать левые интервики. Мастер теней (master of shadows), 14:56, 22 May 2016 (UTC)
Код, который достаёт ОКТМО и без того излишне сложный, ещё и этим его нагружать не стоит. Проще как сейчас - импортировали данные, увидели, что наблюдаются конфликты, исправили проблемные элементы. — Ivan A. Krestinin (talk) 14:59, 22 May 2016 (UTC)
    • Вроде основную массу разгрёб, осталось чуть-чуть для ручного разбора: [17] и [18], хотя наверное можно найти и ещё. — Ivan A. Krestinin (talk) 14:54, 22 May 2016 (UTC)
Хорошо. Мастер теней (master of shadows), 14:56, 22 May 2016 (UTC)
Вот все НП, привязанные к рекам: [19]. Мастер теней (master of shadows), 20:16, 22 May 2016 (UTC)
Сократил этот список до трёх элементов. — Ivan A. Krestinin (talk) 20:34, 22 May 2016 (UTC)

Wrong P17 on bot edit[edit]

This is wrong https://www.wikidata.org/w/index.php?title=Q464980&diff=prev&oldid=328852199Finn Årup Nielsen (fnielsen) (talk) 20:34, 11 May 2016 (UTC)

Литва[edit]

Добрый день! Можете потянуть данные из литовских lt:Šablonas:Ltgyv и lt:Šablonas:Seniūnija? --Чаховіч Уладзіслаў (talk) 17:22, 14 May 2016 (UTC)

Да, всегда Литва. Сенюнии (староства) - нижний уровень адм. деления Литвы, поэтому импортировать нужно sen, хотя, возможно, некоторые населенные пункты имеют районный статус, т.е. нет sen, только sav. Этих свойств вполне достаточно. --Чаховіч Уладзіслаў (talk) 07:47, 15 May 2016 (UTC)
Староства импортировал вместе с country (P17), located in the administrative territorial entity (P131), instance of (P31). А вот с населёнными пунктами не совсем хорошо получается, для многих НП в качестве located in the administrative territorial entity (P131) уже указан район и непонятно: не то это ошибка и нужно исправлять на староство, не то этот НП действительно районного подчинения. Буду импортировать без located in the administrative territorial entity (P131). — Ivan A. Krestinin (talk) 14:20, 15 May 2016 (UTC)

Отчёты[edit]

Приветствую.

Предлагаю заменить в списках точки на нумерацию.

К примеру, Wikidata:Database reports/Constraint violations/P884 — если посещал страницы, фиг поймёшь, какие только что исправлял, а какие нет. С номерами будет попроще. Мастер теней (master of shadows), 11:01, 24 May 2016 (UTC)

  • Доброго вечера, в принципе можно, но всё равно будет не вполне удобно. Возможно при разгребании удобнее использовать SPARQL запросы, которые есть в шаблонах ограничений, их результаты обновляются довольно быстро. Правда иногда они возвращают ерунду. — Ivan A. Krestinin (talk) 18:45, 24 May 2016 (UTC)

imported from:Tajik Wikipedia[edit]

Do you remember where this comes from? I have seen some of these statements now, and they have had in common that they were about something in Brazil and they had Tajik Wikipedia as source, but the claims stated something (very) different. -- Innocent bystander (talk) 06:50, 30 May 2016 (UTC)

  • It is looked like error in tgwiki, fixed. — Ivan A. Krestinin (talk) 07:42, 30 May 2016 (UTC)
    Thanks! Unfortunatly, I do not remember where I found the other statements (which then changed, but not noted you since I though it was a single-case-mistake). I know I visited tgwiki, and since I read maps of South America better than I read Tajik text, I trusted the page to be correct. -- Innocent bystander (talk) 15:09, 30 May 2016 (UTC)

GNDs imported from VIAF[edit]

Hi Ivan, KrBot seems to have stopped to correct format violations of GNDs that have been imported from VIAF, see: Wikidata:Database reports/Constraint violations/P227. There are to many erros to correct them by hand. Can you restart the bot? --Kolja21 (talk) 14:04, 4 June 2016 (UTC)

Hello, GND site request format was changed. Fixed. — Ivan A. Krestinin (talk) 08:29, 5 June 2016 (UTC)
Great. Thank you! --Kolja21 (talk) 10:13, 5 June 2016 (UTC)
Hi Ivan, thanks a lot, indeed. In the GND ID (P227) case there are many manual re-imports of the unsuitable GND identifiers from VIAF, so they also show up as duplicate values in the report and then vanish again after your bot has taken care of their form and then removed them as duplicate value. Even when everything goes smooth this takes about four days until all reports are neutral again. Would it therefore be possible to couple your bot's repair actions of "number is formally wrong but I know how to fix" with the much more frequent updates on Wikidata:Database reports/Constraint violations/Mandatory constraints/Violations ? -- Gymel (talk) 05:59, 7 June 2016 (UTC)

Wrong VIAF on Per Boye Hansen (Q2069926)[edit]

The bot sets Per Boye Hansen (Q2069926) to a wrong VIAF (85219782) [20] Per Boye Hansen is not the same as Per Christian Hansen. — Finn Årup Nielsen (fnielsen) (talk) 08:30, 6 June 2016 (UTC)

The identifier was imported from VIAF. This was single pass import. So you can just delete wrong identifier. Bot will not re-import it. — Ivan A. Krestinin (talk) 17:14, 6 June 2016 (UTC)

"Single value" violations[edit]

Hi Ivan, Wikidata:Database reports/Constraint violations/P227 has now more than 1.000 "single value" violations. If GND ID (P227) is used with a qualifier like pseudonym (P742) the violation is o.k. (= no error). Is there a way your bot can list violations with qualifiers separately? --Kolja21 (talk) 11:44, 11 June 2016 (UTC)

  • Hello, just specify qualifiers in "separators" parameter. Example: [21]. — Ivan A. Krestinin (talk) 21:07, 11 June 2016 (UTC)
    • Great. I'll give it a try: [22] --Kolja21 (talk) 00:23, 12 June 2016 (UTC)

Constraint:Contemporary[edit]

Hi, Ivan.

Would you like to share your comments about a proposal for a new constraint? If you don't have the time, you can simply ignore this message.

Thanks in advance, and best regards, --abián 20:15, 9 July 2016 (UTC)

Please, feel free to tell me if I can help with this issue in some way. --abián 16:40, 15 July 2016 (UTC)
We need to close this issue and distinguish every {{Constraint:Diff within range}} from the rest here. Anyway, I'm still thinking that the discussion page and the constraint violations page get too long and complex by this system (the good point is that I've discovered a new bug thanks to that xD), and a {{Constraint:Contemporary}} would significantly improve this point if it was included in link properties (A-B-C), where it would appear only once and no parameters would be needed. --abián 11:31, 23 July 2016 (UTC)
Hi again. I understand you don't have the time to help with this, and I would like to help you, but I cannot find your bot's code. What can we do? --abián 10:14, 29 July 2016 (UTC)
Hello, you are right, I have no enough time to implement this now. Bot`s code is not published yet. But you can create separate report for now. — Ivan A. Krestinin (talk) 16:28, 29 July 2016 (UTC)
I fear that managing a separate report system that works in a completely different way wouldn't be the best idea. I think it's important for Wikidata that you publish KrBot's code as soon as possible so that the community can help you and that we can prevent any possible data loss. KrBot is too important for all of us! :-) --abián 20:35, 25 September 2016 (UTC)

Category:Pages using invalid self-closed HTML tags[edit]

Your bot currently generates on the constraint pages HTML of the form <span />. Can you please change this behavior either to use {{Anchor}} or to use <span></span>? --Izno (talk) 13:53, 15 July 2016 (UTC)

  • ✓ Done, please wait for the next reports update. — Ivan A. Krestinin (talk) 16:03, 16 July 2016 (UTC)

One of constraint on P17[edit]

Hey. Q55 is in the list of the "one of" constraint on Property talk:P17. But on WD:Database reports/Constraint violations/P17#One of all items with P17=Q55 are listed as violations. --Pasleim (talk) 09:13, 21 July 2016 (UTC)

Incorrect constraint violations list[edit]

this list generated by your bot does not seem to be correct. According to queries to check this constrain on Property talk:P1472, for last few days, there are ~5 pages that violate that constrain. --Jarekt (talk) 12:41, 27 July 2016 (UTC)

  • It is looked strange. Sometimes Wikidata dumps miss some changes. Try to make any change of the item and wait for two days. — Ivan A. Krestinin (talk) 16:32, 29 July 2016 (UTC)

PetScan[edit]

Ivan, could you translate for PetScan for me 'P579 and P2695'? thx --Chris.urs-o (talk) 11:08, 2 August 2016 (UTC)

  • Hello, sorry for delayed response. I am not fully understand your question. PetScan is oriented to category scanning. Request "P579 and P2695" can be queried using WDQ. — Ivan A. Krestinin (talk) 14:58, 4 August 2016 (UTC)
Thx. AutoList 2 is off and I can not write PetScan or WDQ. But thank you very much for the programm. Regards --Chris.urs-o (talk) 16:21, 4 August 2016 (UTC)
@Chris.urs-o: you can make WDQ requests here, if you need some. --Edgars2007 (talk) 16:49, 4 August 2016 (UTC)
Thx, sry --Chris.urs-o (talk) 19:02, 4 August 2016 (UTC)

Whitelists for constraint violations[edit]

Regarding Wikidata:Database reports/Constraint violations/P2880, would it be possible for some single value violations to be whitelisted as valid? Sometimes an article in the database has multiple entries for some reason, so having multiple values is valid. Marking off the valid ones would help make sure I don't repeatedly look at them while screening the ones that genuinely are constraint violations. Thanks, James Hare (NIOSH) (talk) 16:34, 4 August 2016 (UTC)

@James Hare (NIOSH), Harej: that is possible with |exceptions=. See for example Property talk:P2593. Note, that then you have to remove |mandatory=true, if it's included in constraint. --Edgars2007 (talk) 16:47, 4 August 2016 (UTC)
Edgars2007, that should work. Thank you! James Hare (NIOSH) (talk) 20:19, 4 August 2016 (UTC)

Названия НП-Беларусь[edit]

Добрый день! Можете обновить белорусские названия населенных пунктов Беларуси в Викиданных, использовав для этого названия из карточки НП-Беларусь, а те названия, которые поселения имеют сейчас в Википедии и Викиданных (а они разные), перекинуть в дополнительные варианты? А то Вёска NNN, NNN сельсавет выглядит в карточках жутко, а Аграгарадок NNN в карточке персон, рожденных, например, в 19 в. - нелепо. --Чаховіч Уладзіслаў (talk) 07:04, 7 August 2016 (UTC)

  • Здравствуйте, это уже не вполне тривиально, т. к. названия каких-то НП участники могли задать руками и весьма вероятно, что будет некорректным менять его ботом на какой-то ещё вариант. Но если вы можете как-то сформулировать правило по которому бот мог бы выделить "плохие" названия, то их я могу поменять. — Ivan A. Krestinin (talk) 08:32, 7 August 2016 (UTC)
По-идее участники должны были задавать руками то же название, что и в карточке статьи, если же они задавали названия до заливки населенных пунктов в Белорусской Википедии, то вполне вероятно, что такие названия могут быть ошибочными, несогласованными с нормативным справочником названий, по которому велась заливка. В любом случае однозначно необходимо исправить названия ключами Вёска, Хутар, Горад, Гарадскі пасёлак, Аграгарадок, а также все названия с запятыми. --Чаховіч Уладзіслаў (talk) 08:56, 7 August 2016 (UTC)

Identical birth dates (no P570)[edit]

Hi Ivan,

Wikidata:Database reports/identical birth and death dates seems to work quite well. I wonder if you would want to expand it to living people: Wikidata:Database reports/identical birth dates (living people)/1

On Wikidata talk:Database reports/identical birth and death dates/1, there is a SPARQL query that found some.

Maybe as exclusion we could attempt to skip those that have different nationalities (P27) and occupations (P106). So adding these properties would get them removed from the list.
--- Jura 08:47, 7 August 2016 (UTC)

The following query uses these:

SELECT DISTINCT ?item1 ?item2  ?d1
{
  	?item1 wdt:P569 ?d1 .
  	FILTER (?d1 > "1980-01-01T00:00:00Z"^^xsd:dateTime && ?d1 < "1985-01-01T00:00:00Z"^^xsd:dateTime) .
  	OPTIONAL { ?item1 wdt:P570 ?dummy1 } FILTER ( !bound(?dummy1) ) .
  	?item2 wdt:P569 ?d1 .
  	OPTIONAL { ?item2 wdt:P570 ?dummy2 } FILTER ( !bound(?dummy2) ) . 
  	?item1 wdt:P31 wd:Q5 .
  	?item2 wdt:P31 wd:Q5 .
  	?item1 wdt:P21 ?g .
  	?item2 wdt:P21 ?g .
	OPTIONAL { ?item1 wdt:P27 ?l1 }		OPTIONAL { ?item2 wdt:P27 ?l2 }		FILTER ( !( BOUND(?l1) && BOUND(?l2) && ?l1 != ?l2) )
	OPTIONAL { ?item1 wdt:P106 ?n1}		OPTIONAL { ?item2 wdt:P106 ?n2}		FILTER ( !( BOUND(?n1) && BOUND(?n2) && ?n1 != ?n2) )
  	OPTIONAL { ?item1 wdt:P19 ?b1 }		OPTIONAL { ?item2 wdt:P19 ?b2 }		FILTER ( !( BOUND(?b1) && BOUND(?b2) && ?b1 != ?b2) )
  	OPTIONAL { ?item1 wdt:P22 ?f1 }		OPTIONAL { ?item2 wdt:P22 ?f2 }		FILTER ( !( BOUND(?f1) && BOUND(?f2) && ?f1 != ?f2) )
  	OPTIONAL { ?item1 wdt:P25 ?m1 }		OPTIONAL { ?item2 wdt:P25 ?m2 }		FILTER ( !( BOUND(?m1) && BOUND(?m2) && ?m1 != ?m2) )
  	OPTIONAL { ?item1 wdt:P646 ?z1 }	OPTIONAL { ?item2 wdt:P646 ?z2 }	FILTER ( !( BOUND(?z1) && BOUND(?z2) && ?z1 != ?z2) )
  	OPTIONAL { ?item1 wdt:P413 ?w1 }	OPTIONAL { ?item2 wdt:P413 ?w2 }	FILTER ( !( BOUND(?w1) && BOUND(?w2) && ?w1 != ?w2) )  
  	FILTER ( str(?item1) < str(?item2)  )
  	MINUS { ?item1 wdt:P7 ?item2 }		MINUS { ?item1 wdt:P9 ?item2 }		# twins
  	MINUS { ?item1 wdt:P1889 ?item2 }	MINUS { ?item1 wdt:P460 ?item2 }	# identified issues
}
LIMIT 50

Try it!


I did some tests with SPARQL (see above). It seems that this could generate a lot of items. Items that have nothing more than P569/P21 turn out a lot of results. Maybe this is suitable for a game instead: add a statement to each item with a property that differentiates them.
--- Jura 08:23, 17 August 2016 (UTC)
  • It gives reasonable results when checking for identical labels: User:Jura1/test32. SPARQL times out if more than one 1 year is done at once. For the other approach, User:Jura1/test31 looks interesting, but would need some interface.
    --- Jura 09:24, 17 August 2016 (UTC)

Constraint violation summary[edit]

Hi! Maybe it would be possible to create something like one constraint violation summary page for all props? Section for P1146 it would return (in table format or list format):

  • P1146 - "Format" violations - 0
  • P1146 - "Unique value" violations - 0
  • P1146 - "Single value" violations - 18 (+ 2 exceptions)
  • etc.

It would be probably useful to know, which props and constraint violations has the biggest problems. I'm not saying, that I would personally be very active, but if it's pretty trivial to create such report (and I don't see anything very hard there, but I may be wrong), then why not? --Edgars2007 (talk) 09:32, 13 August 2016 (UTC)

Update frequency of Wikidata:Database reports/Constraint violations/Mandatory constraints/Violations[edit]

It seems to take very long to get this page updated every day, can you explain? I would love to clean some constraints violations, but it is full of old values. Sjoerd de Bruin (talk) 19:55, 16 August 2016 (UTC)

  • Bot removes fixed values hourly. But sometimes bot is busy by another long tasks. For example full constraint violations reports update takes 6 hours. — Ivan A. Krestinin (talk) 21:44, 16 August 2016 (UTC)
  • The page has the size of some months ago again, hourly updates again. \o/ Of course, it would be great to hit zero. But this will make it go easier. Sjoerd de Bruin (talk) 11:28, 29 August 2016 (UTC)

Wikidata:Database reports/Constraint violations/P625[edit]

There is a timeout at Wikidata:Database reports/Constraint violations/P625. I already proposed, that single value constraint may be omitted or limited, that can ease the calculation. --Jklamo (talk) 08:27, 17 August 2016 (UTC)

  • Hello, it was HTTP timeout, not data calculation timeout. Some network issue happened. — Ivan A. Krestinin (talk) 17:09, 17 August 2016 (UTC)
    • Can't we move to WMF Labs clusters? I assume, that would resolve some problems (also for section above). --Edgars2007 (talk) 18:52, 17 August 2016 (UTC)

Constraint violations report[edit]

Добрый день! Иван, не могли бы вы в отчётах "Single value" violations около каждого из значений показывать его ранг? Как правило в таких случаях, что показывать редактор указывает именно рангом. См., например, no label (Q25939342) попавший в Wikidata:Database reports/Constraint violations/P1973. -- Sergey kudryavtsev (talk) 07:00, 27 August 2016 (UTC)

PS: Можно использовать иконки вроде s:ru:Файл:Statement prefered rank icon.gif, s:ru:Файл:Statement normal rank icon.gif и s:ru:Файл:Statement not recommended rank icon.gif. -- Sergey kudryavtsev (talk) 07:08, 27 August 2016 (UTC)

  • Добрый день, сейчас бот не зачитывает ранги совсем, сделать конечно можно, но определённые трудности есть. Кстати можно немного по-другому поступить: для устаревшего ID указать какой-нибудь квалификатор, вроде "дата деактуализации" и это свойство указать в параметре "separators" шаблона {{Constraint:Single value}}. Ещё, как вариант, можно не пытаться коллекционировать устаревшие и невалидные идентификаторы, а просто удалять их, так делается для VIAF, GND. — Ivan A. Krestinin (talk) 11:48, 28 August 2016 (UTC)
    А что за параметр "separators"? Как он действует? -- Sergey kudryavtsev (talk) 03:54, 30 August 2016 (UTC)
    Всё не хватает времени задокументировать его. В этом параметре перечисляются свойства-квалификаторы. Если в каком-то элементе два значения, но они отличаются одним из указанных квалификаторов, то ограничение {{Constraint:Single value}} пропустит такой элемент. — Ivan A. Krestinin (talk) 20:14, 1 September 2016 (UTC)
    Да, полезная возможность. А отсутствие квалификатора, перечисленного в separators будет считаться особым значением, отличным других? Судя по Wikidata:Database reports/Constraint violations/P947#Single value это не так, поэтому я добавил там в separators ещё и excluding (P1011), посмотрим как ваш бот отработает это ограничение. -- Sergey kudryavtsev (talk) 08:58, 15 September 2016 (UTC)

Вы не могли бы заставить бота обновить Wikidata:Database reports/Constraint violations/P947, а то он последний раз несколько месяцев назад обновлялся (2016-07-17)... -- Sergey kudryavtsev (talk) 11:45, 23 September 2016 (UTC)

Спасибо, что обновили этот отчёт, но похоже, что ваш бот не воспринял параметр |separators=P407, P1011, который вы рекомендавали выше. Например, в отчёт попал Galileo Galilei (Q307), у которого значения отличаются значениями квалификаторов. -- Sergey kudryavtsev (talk) 17:40, 24 September 2016 (UTC)
Оба значения должны иметь квалификатор одного типа и значения этого квалификатора должны отличаться. Возможно в данном случае правильнее будет вместо задания квалификатора P1011, создать специальный элемент с заголовком "все языки, кроме русского" и выставлять его в качестве значения свойства P407. — Ivan A. Krestinin (talk) 12:07, 25 September 2016 (UTC)
По-моему отсутствие квалификатора должно считаться отдельным значением, отличным ото всех других (как значение NULL в базах данных). Или, если это сделать технически сложно, значением somevalue (неизвестно). А сейчас бот работает так, будто отсутствие квалификатора — это значение, равное любому другому. -- Sergey kudryavtsev (talk) 18:49, 27 September 2016 (UTC)

Units redir[edit]

Hi Ivan. Q15639371 was redirected to Q13147228, but is still used in units (538 times). Would you consider adding this to your bot or shall I ask at WD:Bot requests?
--- Jura 16:40, 29 August 2016 (UTC)

KrBot - handling unknown values[edit]

Hi again, I have a question about KrBot's reports. For Slovene Cultural Heritage Register ID (P1587) (report here), I implemented a change, entering heritage status (P1435) of some objects with the system setting unknown value. This is on purpose, the registry actually has some objects that have officialy "unknown status" as monuments. The question is, what to put on the property's description page so that the bot will label this as an accepted value? As you can see in the report now, it labels them as violations and "somevalue". — Yerpo Eh? 07:32, 2 September 2016 (UTC)

Thanks, this works. The label "somevalue" is a little confusing, but I guess it's not so easy to modify that. — Yerpo Eh? 12:31, 9 September 2016 (UTC)

Value types[edit]

Hi, somehow the value type violations on P3096 seem odd. There is a similar report on project chat. You might want to look into it.
--- Jura 04:07, 9 September 2016 (UTC)

Something went wrong with dumps?[edit]

Hi! Yesterday the bot made such update, lowering number of items, that has this prop. Excluded was also for example Luisa Mercado (Q24701033), that hasn't been changed for a while. And today the number of items hasn't been restored to approx. previos number, so I assume this problem is still actual. This prop of course isn't the only one, there was some weirdeness on other covi pages, too. --Edgars2007 (talk) 05:11, 9 September 2016 (UTC)

Удаление прежних кодов ОКТМО[edit]

Добрый вечер, Иван! Не могли бы пояснить, почему вы удаляете устаревшие коды ОКТМО? В пояснении написано, что Marking erroneous statements as deprecated instead of simply deleting such statements has three benefits. Игорь Темиров (talk) 19:50, 9 September 2016 (UTC)

  • На Викиданных большой бардак творится с кодами ОКТМО, никак не удаётся навести порядок в актуальных кодах, в этих условиях поддерживать ещё и устаревшие коды нет никаких сил, проще просто удалять, как делается для многих других свойств-идентификаторов. — Ivan A. Krestinin (talk) 07:55, 11 September 2016 (UTC)

Wikidata:Database reports/Constraint violations/P625 [2][edit]

Hi, there was no update of that constraint since 11 September. It may be related to addition of constraint (here). --Jklamo (talk) 07:13, 14 September 2016 (UTC)

If I'm the cause, then I'm sorry and we can revert that edit of mine. Didn't intentionally broke update :) But isn't there some way to generate value statistics table for P625? At least, a one-time update somewhere else, not on covi page. I think it would really be interesting to see, what the situation is. --Edgars2007 (talk) 08:03, 14 September 2016 (UTC)
Maybe not, it just mine hypothesis. I think your table may be crated using query, maybe this is what you looking for (autolist query, as I am still not familiar enough with sparql). --Jklamo (talk) 11:38, 14 September 2016 (UTC)
No, using WDQ is not possible to get the table of value statistics. changed the header, so one can get to right section --Edgars2007 (talk) 15:26, 14 September 2016 (UTC)
The report is too large now. I fixed error processing code to highlight the issue. But I need to think about working with too large report. — Ivan A. Krestinin (talk) 12:25, 25 September 2016 (UTC)

SetReference[edit]

При изменении ОКТМО добавление ботом нового ОКТМО происходит без проблем, а вот при добавлении к нему ссылки пишет, что такого claim нет. Должно пройти какое-то время (не засекал, но не меньше 10 минут) прежде чем он даёт добавить добавить ссылку. У вас это происходит одномоментно. Не подскажете, в чём здесь загвоздка? Игорь Темиров (talk) 05:18, 15 September 2016 (UTC)

  • Не знаю, впрочем я уже какое-то время не редактировал ничего, возможно с каким-нибудь очередным обновлением у меня тоже будут проблемы. — Ivan A. Krestinin (talk) 12:09, 25 September 2016 (UTC)

Wikidata:Database reports/Constraint violations/P1792[edit]

The "target required claim" reports on Wikidata:Database reports/Constraint violations/P1792 seem to be wrong. Basically all items with a category of associated people (P1792) claim are listed. Strangely, on Wikidata:Database reports/Constraint violations/Mandatory constraints/Violations the report length for P1792 is shorter but there still many false positives. --Pasleim (talk) 23:56, 20 September 2016 (UTC)

  • ✓ Done, please wait for the next report update. — Ivan A. Krestinin (talk) 12:21, 25 September 2016 (UTC)
Thanks. --Pasleim (talk) 14:42, 25 September 2016 (UTC)

KrBot changing halfbrother to martial half-brother[edit]

See the discussion in the Project Chat: https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat&action=view&section=37 ChristianKl (talk) 16:26, 27 September 2016 (UTC)

Update stuck?[edit]

Identical dates hasn't been updated.
--- Jura 09:23, 29 September 2016 (UTC)

Bot had some troubles with Q16086943. Now all is ok. It is difficult to identify the issue now. — Ivan A. Krestinin (talk) 10:23, 2 October 2016 (UTC)
Three items that became one? I don't think I have had that before. Thanks for having looked into it.
--- Jura 10:27, 2 October 2016 (UTC)
Is it possible that the main constraints reports have gotten so long to run that it just doesn't get to update this? Is there something we can do to limit the main constraint reports? Maybe for some of the constraints, it's not necessary to compute the violation and leave it to Special:ConstraintReport?
--- Jura 08:02, 6 October 2016 (UTC)
Would you check it once more?
--- Jura 15:33, 11 October 2016 (UTC)
Fixed, constraint report task took too much time. — Ivan A. Krestinin (talk) 13:50, 15 October 2016 (UTC)

Database Constraint report for 3217[edit]

Should I do something to get the constraint violation report starting to run? - Salgo60 (talk) 13:37, 4 October 2016 (UTC)

No, you did everything you had to. That page will be updated in next 24 h. --Edgars2007 (talk) 13:42, 4 October 2016 (UTC)

Single value constraint[edit]

For some identifiers we had deprecated values. Some contributors like to add these old data but this is a problem for the constraint Single value often used with identifiers. Do you know if it is possible to consider only value with a normal rank when perform the Single value check ? Thanks Snipre (talk) 07:28, 24 October 2016 (UTC)

Or rather, not to consider deprecated rank (since preferred rank is also possible @Snipre:). --Izno (talk) 12:01, 24 October 2016 (UTC)
@Izno: If we assume that an identifier is unique then we can't choose between several values so the preferred rank should not be used. Again the single value constraint implies that you have only one valid value at a certain date. Using preferred rank and deprecated rank only is not correct in this case. Preferred rank should be used only when you have several normal rank values. This is not the case here. Snipre (talk) 12:15, 24 October 2016 (UTC)
@Snipre: My point is that only "deprecated" should be excluded if there are multiple claims with a single value constraint. In other words, "preferred" should also produce a constraint error, if there are normal ranks also present. --Izno (talk) 12:18, 24 October 2016 (UTC)
@Izno: Ok with you if your "possible" is a possibility but not the acceptance as a normal case. Snipre (talk) 14:33, 24 October 2016 (UTC)
Current bot implementation ignores ranks at all. You can use parameter "separators=start time (P580), end time (P582)". Collecting deprecated identifiers is not looked as very good idea. The identifiers can be unverifiable. Are you know data client who need this information? Deprecated values are need to be processed in all data clients. Are you want to import all deprecated values to Wikidata? — Ivan A. Krestinin (talk) 20:46, 24 October 2016 (UTC)
The objective is first to know if it is possible and what are the technical consequences. We started a discussion in Project Chemistry about one identifier, the CAS number. We don't have a lot of deprecated values per item so it doesn't concern hundred thousands values.
The problem of the source is not a real problem: most of databases about chemicals using CAS numbers already implement some kind of redirection for deprecated values because CAS number is a worldwide identifier used in research as well than in the chemicals trading. So most of use still keep a trace of old values. Then the idea is not to import all possible deprecated values for CAS number because the original database is a proprietary database. So we have access only to the CAS number of the chemicals which the most use in the economy.
This idea is not supported by 100% of the contributors who took part to the discussion so the question is only technical. Snipre (talk) 19:41, 26 October 2016 (UTC)

Bn description[edit]

Hi, Currently your bot (KrBot) automatically adds description ("Wikimedia category") in 8 languages. Is it possible to add bengali description ("উইকিমিডিয়া বিষয়শ্রেণী") also? --Aftabuzzaman (talk) 15:57, 26 October 2016 (UTC)

Is it also possible to add hebrew? The description will be ("דף קטגוריה")--Mikey641 (talk) 17:43, 26 October 2016 (UTC)
This algorithm is not very efficient now. My bot adds every description using separate edit. I saw bots that add many descriptions using single edit. Please ask another botmasters. — Ivan A. Krestinin (talk) 19:09, 15 February 2017 (UTC)

Wikidata:Database reports/Constraint violations/P639[edit]

Here your Bot produced an error. Steak (talk) 20:41, 2 November 2016 (UTC)

  • The page contains too many violations. Constraints are need to be reviewed to identify the most violated constraint. — Ivan A. Krestinin (talk) 19:41, 18 November 2016 (UTC)

External use bot[edit]

Hi, Ivan. I noticed that your bot is adding to "external use" template usage on ruwiki. I appreciate that as it is important to have that template as complete as possible. Do you have a script for that and it is possible to extend that to other wikis? --Jklamo (talk) 08:54, 3 November 2016 (UTC)

+1 to Jklamo request. It would be great to have a similar bot (or as script) to run also in hewiki. Eran (talk) 19:17, 14 November 2016 (UTC)
The code is dump-based and resource-consuming. It is not simple to scale it to all wikies. Do you know some service or API that allows to search {{#property... like constructions in all wiki templates? — Ivan A. Krestinin (talk) 19:45, 18 November 2016 (UTC)
Why not mw:API:Search? There you can search also by regex. --Edgars2007 (talk) 02:41, 25 December 2016 (UTC)
@Edgars2007: Sorry, I miss your post. mw:API:Search does not allow to search in page wiki-text. It searches in generated text. Try to find {{#property: construction. — Ivan A. Krestinin (talk) 12:31, 15 January 2017 (UTC)
There is insource: command for that, I thought you know about it, that's why I didn't mention it. See this example in API sandbox. --Edgars2007 (talk) 12:45, 15 January 2017 (UTC)
I think that https://quarry.wmflabs.org is able to search specific text in wiki templates.--Jklamo (talk) 12:43, 15 January 2017 (UTC)
No, page wikitext isn't stored in SQL databases (that means, it's not available in quarry.wmflabs.org). --Edgars2007 (talk) 12:45, 15 January 2017 (UTC)
Thank you for "insource" keyword. Bot processed the most projects. Please notify me if bot miss something. — Ivan A. Krestinin (talk) 12:41, 23 January 2017 (UTC)
The thing bot "missed" is using other constructions than #property: to get data. You come from ruwiki, which is an excellent example. And then there are those Lua modules :) Much fun. Just loud thinking, no complains. --Edgars2007 (talk) 08:48, 26 January 2017 (UTC)
Also not complain, but profound thanks. What about deleted properties, did you process them as well? If not, I am very interested at least in no label (P7) and no label (P9).--Jklamo (talk) 09:10, 6 February 2017 (UTC)
Please provide examples of missed usages. Bot processed deleted properties too. But it has troubles with {{ExternalUse}} adding. Bot does not recreate property talk pages. As for P7/P9 bot failed to find appropriate place to the template. I added it manually and will restart bot. — Ivan A. Krestinin (talk) 09:37, 6 February 2017 (UTC)

Why?[edit]

Why did you do this edit? /ℇsquilo 10:21, 7 November 2016 (UTC)

Because the claim is exact dublicate of another claim for this property at that item. Read the edit summary. --Edgars2007 (talk) 10:40, 7 November 2016 (UTC)

Constraint violation reports[edit]

Your usually helpful bot hasn't updated any of the Wikidata:Database reports/Constraint violations/P* pages since 5 November. Is there an issue? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:21, 9 November 2016 (UTC)

Just adding myself to the list of those who miss constraint violations updates.--Jklamo (talk) 00:22, 14 November 2016 (UTC)
+1. Lymantria (talk) 14:29, 14 November 2016 (UTC)
It's a pity that wikidata depends in a critical functionality on one user. Why cannot the Foundation run a reliable bot? Steak (talk) 20:07, 14 November 2016 (UTC)
Really, Steak? This outage is unusual and I hope Ivan is well. --Succu (talk) 21:34, 14 November 2016 (UTC)
Steak! Most (but not all) constraint reports can be seen if you follow the SPARQL-link in the constraints templates. Having a set of at least two bots for the most critical reports is not a bad idea at all. -- Innocent bystander (talk) 11:58, 15 November 2016 (UTC)
Not all reports are "SPARQLed" and there is anther added value of daily reports - if you are checking reports daily there is possibility to compare daily reports and detect bad imports or newbie users edits. --Jklamo (talk) 13:15, 15 November 2016 (UTC)
I had bad troubles with hardware. Now all is fine and bot will update the reports during next 5-10 hours. Pasleim reimplement the bot as I see. So the system is reserved now. Thank you, Pasleim. — Ivan A. Krestinin (talk) 19:50, 18 November 2016 (UTC)
That are good news. Thanks a lot. --Succu (talk) 20:55, 18 November 2016 (UTC)
Thanks much, Ivan and Pasleim!! Lymantria (talk) 10:07, 20 November 2016 (UTC)
Good to hear. Thank you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:10, 20 November 2016 (UTC)

Thank you[edit]

Hello, this is amazing. I made a mistake and am unable to fix it using tools that I can control. There are identical problems in most Czech municipalities. Could you run through all of them and do the same? Thank you, --Vojtěch Dostál (talk) 14:16, 16 November 2016 (UTC)

Could you provide some unfixed item? — Ivan A. Krestinin (talk) 19:52, 18 November 2016 (UTC)

Publish the source of KRBot please[edit]

Hi Ivan, you're running a bot without it's source published. You vanished and the bot broke down so nobody could recover it. Please publish your source somewhere (https://github.com/?) as soon as possible. No matter that it might be dirty, clean up needed, etc etc. Just dump online somewhere and add the link to User:KrBot. Pasleim now spend(/wasted) a lot of time on trying to recover the service, I would hate to see that happen again. Multichill (talk) 10:50, 21 November 2016 (UTC)

Hi Ivan, it's been a month since I posted this message. It would be nice if you could respond. Multichill (talk) 09:14, 21 December 2016 (UTC)
It looks like the response is no. I think similar code is already in Mediawiki (Special:Constraint report), it's just that WMDE doesn't want to work on it.
--- Jura 11:24, 21 December 2016 (UTC)
Let's not do assumptions and have Ivan respond. Multichill (talk) 11:29, 21 December 2016 (UTC)
The code is written on C++ and uses custom libraries. It does not have any documentation. It has unused parts, for example for property-described constants. The reports update algorithm is resource consuming. It eats 32 Gb of operating memory. Do you really need this code? And I will publish the code in case of my exit from the project of course. — Ivan A. Krestinin (talk) 18:13, 21 December 2016 (UTC)
Sounds like you forked wdq. ;-)
The current Bus factor (Q1812883) is one so yes, you should definitely publish it somewhere.
I don't want you to leave, but if you do decide to take a long holiday, at least the code is available to fork. It's also nice for other people who are re-implementing parts to be able to have a look at existing code how it's solved there. Multichill (talk) 22:39, 21 December 2016 (UTC)

Deception[edit]

If you look at this edit, everything seems fine. But what you don't see in the difference is that the bot moved one claim from a slot that had no references into a slot that had two references. The edit therefore implies that the newly moved value is the one with references. This is deceptive, as the new claim value was not the one to which those references apply. --EncycloPetey (talk) 18:50, 11 December 2016 (UTC)

Identifier 179917357 is equal to 101760867. Both values identify single record in VIAF database. So reference applied to one value can be applied to the second value. — Ivan A. Krestinin (talk) 12:47, 12 December 2016 (UTC)

Unmerge[edit]

Откатите, пожалуйста, перенаправления типа этого. --Infovarius (talk) 21:28, 20 December 2016 (UTC)

И эти. --Infovarius (talk) 21:40, 20 December 2016 (UTC)
✓ DoneIvan A. Krestinin (talk) 04:35, 21 December 2016 (UTC)
Может, где-то есть страница запросов? :) Вот ещё: [23]. --Infovarius (talk) 13:00, 26 December 2016 (UTC)
✓ DoneIvan A. Krestinin (talk) 05:13, 27 December 2016 (UTC)

Batch rollback request[edit]

Your bot changed a large number of instance of (P31)  MediaWiki configuration settings (Q21481290) to instance of (P31)  Manuel Candamo (Q170059), where it doesn't know the merge action of Q21481290 is invalid. --Liuxinyu970226 (talk) 01:03, 24 December 2016 (UTC)

Would you do the same for Gottfried (Q16746939) > no label (Q3566065). Thanks.
--- Jura 14:56, 26 December 2016 (UTC)

✓ DoneIvan A. Krestinin (talk) 05:13, 27 December 2016 (UTC)

Change the external link for CAS number[edit]

Hi. Just to inform you that the tool that lists all URL with a CAS number in order to access external databases is dead. Can you please please the links in Wikidata:Database_reports/Constraint_violations/P231 by https://chem.nlm.nih.gov/chemidplus/rn/* where * is the CAS number ? So https://tools.wmflabs.org/magnustools/cas.php?cas=63800-37-3 will become https://chem.nlm.nih.gov/chemidplus/rn/63800-37-3. Thank you. Snipre (talk) 20:36, 9 January 2017 (UTC)

@Snipre: the only thing you have to do is change formatter URL (P1630) at CAS Registry Number (P231) ($ there is kind of the same thing as * in your example here) --Edgars2007 (talk) 06:06, 10 January 2017 (UTC)
Edgars2007 Thank you. Snipre (talk) 09:07, 10 January 2017 (UTC)

Q27891162[edit]

Regarding Q27891162 I do not think that it is true, that it has a P:366 as Q949149

best regards to your bot, Scoid d (talk) 17:04, 20 January 2017 (UTC)

Bot in hollyday?[edit]

https://www.wikidata.org/w/index.php?title=Wikidata:Database_reports/Constraint_violations/P345&action=history -- MovieFex (talk) 16:45, 25 January 2017 (UTC)

  • Bot worked on long import task. Now it downloads new Wikldata dump. So the report will updated in ~5 hours. — Ivan A. Krestinin (talk) 21:46, 25 January 2017 (UTC)

KrBot {{ExternalUse}}[edit]

Hi. Thanks for maintaining KrBot. It seems that the bot hasn't been indexing property uses on Cantonese Wikipedia. Does it need to be set up separately for each wiki? Deryck Chan (talk) 15:40, 6 February 2017 (UTC)

Hello, bot processes all wikies. But it can miss some unknown constructions. Could you provide links to missed templates? — Ivan A. Krestinin (talk) 15:44, 6 February 2017 (UTC)
It seems that the template w:zh-yue:Template:英格蘭郡明細 (Infobox English county) is only indexed by Property talk:P625 (which ironically doesn't work, because zh-yue has a different implementation of {{coord}} that hasn't been ironed out yet). Your bot's latest batch of edits didn't index that infobox on any of the other properties it uses. Deryck Chan (talk) 11:31, 7 February 2017 (UTC)
I see that the latest run has indexed my template. Is there anything that needs to be done to ensure future uses of Wikidata by zh-yue.wp continues to be done properly? Deryck Chan (talk) 11:06, 8 February 2017 (UTC)
Different projects use different constructions to call Wikidata. Ruwiki uses {{Wikidata|P18}}, zh-yuewiki uses {{#invoke:wikidata|getvalue|P18}}. These constructions are need to be processed separately. So bot can miss some another call types. Just notify me if bot miss something. And I will add new calling convention to bot`s algorithm. — Ivan A. Krestinin (talk) 21:23, 8 February 2017 (UTC)
I wonder if syncing calls of something like w:Template:Uses Wikidata - and its equivalent in other projects - might be a better way of indexing uses of Wikidata. Eventually we'll need a central register of how each wiki calls Wikidata - it will be very useful if KrBot's input dataset (the constructions in each wiki) are listed on an editable wikipage. Deryck Chan (talk) 11:14, 9 February 2017 (UTC)

Rollback request[edit]

The merge from Murakami (Q21494521) to Murakami (Q1953526) was wrong. Could you please undo edits like this? --Pasleim (talk) 10:16, 10 February 2017 (UTC)

The merge from Juri (Q18932166) to Yuri (Q4995000) was wrong. Could you please undo edits like [24]? --Pasleim (talk) 17:04, 18 February 2017 (UTC)

KrBot problem with ExternalUse[edit]

Hi KrBot has duplicated several times ExternalUse for itwiki, examples Property talk:P3109, Property talk:P3513, Property talk:P3442. Could you check your work and fix it? Thanks in advance. --Rotpunkt (talk) 13:50, 11 February 2017 (UTC)

  • Hi, the bug was fixed already. But I did not create algorithm to clean duplicate names. For now you can just revert bot`s edit on these pages. — Ivan A. Krestinin (talk) 17:26, 12 February 2017 (UTC)
Mhm, the problem is in a lot of pages, you should fix your errors, reverting your bot. --Rotpunkt (talk) 20:54, 12 February 2017 (UTC)
Fixed. — Ivan A. Krestinin (talk) 19:05, 15 February 2017 (UTC)

Your feedback matters: Final reminder to take the global Wikimedia survey[edit]

(Sorry to write in Engilsh)

External use[edit]

There is some reason to add no existing template in Externale use like in P106? --ValterVB (talk) 20:06, 24 February 2017 (UTC)

✓ DoneIvan A. Krestinin (talk) 21:54, 24 February 2017 (UTC)

Do not delete the identifier[edit]

Don't delete the identifier IMDb without correcting the entry at the Wiki-Article. There is Wikidata:Database reports/Constraint violations/P345 and it will be fixed, all inclusive. --MovieFex (talk) 02:51, 25 February 2017 (UTC)

Talk pages of constraint reports[edit]

When your bot creates a report page like Wikidata:Database reports/Constraint violations/PNNNN, please can it then create a redirect from the corresponding talk page, Wikidata talk:Database reports/Constraint violations/PNNNN, to the talk page for the property, Property talk:PNNNN, where those constraints are set? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:31, 27 February 2017 (UTC)

Wow, I just saw the redirect creations, that a very good idea! Thank you Pigsonthewing (talkcontribslogs) and Ivan! Cdlt, VIGNERON (talk) 20:51, 1 March 2017 (UTC)

qualifier → reference[edit]

Здравствуйте! Спасибо за помощь,

Thank you for making corrections where I should be using the 'reference' field and not 'qualifier', I agree.

Kind regards, --2015.ww (talk) 05:25, 7 March 2017 (UTC)

Double entry on published in (P1433)[edit]

You bot introduces a double entry on published in (P1433) [25]. — Finn Årup Nielsen (fnielsen) (talk) 13:38, 10 March 2017 (UTC)

Mass revert[edit]

Hi, could you mass revert somehow your bot's "genre: drama" edits? Two items were merged that shouldn't have been. – Máté (talk) 13:24, 17 March 2017 (UTC)

✓ DoneIvan A. Krestinin (talk) 07:24, 18 March 2017 (UTC)

External use[edit]

Hey, based on what did you do that can u do it again for all properties in hewiki? btw: in hewiki we use a template that indicates what properties are used for example he:תבנית:קו רכבת תחתית. thanks--Mikey641 (talk) 11:36, 19 March 2017 (UTC)

Reaction on WD:DNM[edit]

Hi Ivan, I left a message on Wikidata talk:Do not merge where I pinged you. Could you please have a look at it? Q.Zanden questions? 13:44, 21 March 2017 (UTC)

KrBot not updating constraint violation reports[edit]

Hi Ivan, I noticed that KrBot did not update the constraint violation reports yesterday. Is everything working? Cheers. − Pintoch (talk) 09:14, 22 March 2017 (UTC)

Hello, bot started update process some hours ago. It had some troubles, but all is ok now. — Ivan A. Krestinin (talk) 21:05, 22 March 2017 (UTC)

Item labels on constraint report[edit]

Hi there, firstly thanks for running the bot for the constraint report, I've fixed many errors in both wikidata and articles because of it. I've been using Wikidata:Database reports/Constraint violations/P2697 and Wikidata:Database reports/Constraint violations/P2698 and I've noticed that P2698 just has the item numbers (ie Q123456) in the lists, but P2697 shows the item label and number (ie P.G. Wodehouse (Q207515)). Is there a setting somewhere to force the item labels to be shown on all reports? Thanks, The-Pope (talk) 00:45, 23 March 2017 (UTC)

It's because of length of report (if you use {{Q}} too much, you'll get Lua error). Because of relatevily new updates in Module:Wikidata Ivan may reevaluate the threshold, if it's set in bot's code. --Edgars2007 (talk) 13:56, 23 March 2017 (UTC)
Hello, you are right. Current threshold is 450 items. I increase it to 2000. Lets wait the next update. — Ivan A. Krestinin (talk) 18:28, 23 March 2017 (UTC)

Special:Diff/471393970/472494721[edit]

Can't your bot do that in one edit? --Leyo 08:04, 6 April 2017 (UTC)

  • Hello, currently the bot does not have code for making this in single edit. Another bots can do this. — Ivan A. Krestinin (talk) 19:25, 6 April 2017 (UTC)

Three copies of the link[edit]

After this edit, Euripides (Q48305) now has three identical VIAF links to the same target. --EncycloPetey (talk) 23:36, 6 April 2017 (UTC)

Help with a regex for IPv4 routing prefix (P3761)[edit]

Hi again Ivan, Thanks for your fix to the bot last month!

I have tried to fix the regex for the format of IPv4 routing prefix (P3761). It seems to be correct now (as a PCRE), and the web interface accepts well-formed edits, but it looks like KrBot considers that all the statements have an incorrect format. I think the slash was not properly escaped before, which is now fixed. Maybe the regular expressions are cached somewhere and we need to update that cache too? Have a good week-end! − Pintoch (talk) 22:45, 7 April 2017 (UTC)

    • Great, thanks a lot! Actually I did use https://regex101.com and the previous regex did work in that tool (as a PCRE). Funny. − Pintoch (talk) 17:22, 10 April 2017 (UTC)

Updated VIAF[edit]

I have done this in connection with an update of VIAF. The "correct" number was already there. Maybe the bot should check if the number is already there and then maintain the old number and deprecate it? Or do you have any other opinion? — Finn Årup Nielsen (fnielsen) (talk) 17:15, 9 April 2017 (UTC)

Eventually it removes that duplicate.
--- Jura 17:19, 9 April 2017 (UTC)
✓ Done Bot will fix duplicates immediately after appearing. — Ivan A. Krestinin (talk) 05:21, 10 April 2017 (UTC)

Удаление информации[edit]

Чего это бот такое делает: [26] ?ShinePhantom (talk) 03:31, 19 April 2017 (UTC)

В коде бота был баг, большинство правок вызванных этим багом я откатил, но вот эту похоже упустил. — Ivan A. Krestinin (talk) 04:26, 19 April 2017 (UTC)

Удаление Административно-территориальных единиц[edit]

Добрый день! Обратил внимание на несколько правок, сделанных Вашим ботом в части удаления из населённых пунктов и административных образованиях информации об административно-территориальных единицах, как упразднённых, так и ещё вполне действующих. Данная информация имеет значение для викиданных, просьба не удалять её. В частности в русской Википедии от этого зависит корректное заполнение карточек персон, в противном случае получается, что-то вроде Иванов родился в 1781 году в Нижегородской области Российской Федерации. И если можно, прокомментируйте, пожалуйста, правки Вашего бота. Хотелось бы понять масштаб проблемы: это было разовое действие или правка, осуществляемая на регулярной основе? Речь идёт только о России или о других странах тоже? Можно ли её отменить (вернуть как было)? --Ksc~ruwiki (talk) 19:28, 19 April 2017 (UTC) P.S. Прошу прощения, не заметил, что коллега уже затрагивал эту тему, но если можно ответьте на вопросы, чтобы у меня было понимание происходящего. --Ksc~ruwiki (talk) 19:30, 19 April 2017 (UTC)

По поводу элементов, где приведена история принадлежности объекта различным образованиям, где указаны квалификаторы start time (P580)/end time (P582) был баг в коде бота, такие элементы он трогать не должен был. Подобные правки, насколько я вижу, все откатили. Второй случай другой: если упрощённо, то раз Arzamas (Q103789) входит в no label (Q27546988), а no label (Q27546988) входит в Nizhny Novgorod Oblast (Q2246), то нет смысла дополнительно указывать, что Arzamas (Q103789) входит в Nizhny Novgorod Oblast (Q2246). Для России на Викиданных сейчас такая модель используется, про другие государства не знаю. Хотя существуют и другие модели. Например, модель чисто-административного деления. Или модель чисто-территориального деления. — Ivan A. Krestinin (talk) 19:55, 19 April 2017 (UTC)
Смысл есть на самом деле. Дело в том, что no label (Q27546988) сравнительно новое образование и Nizhny Novgorod Oblast (Q2246) для Arzamas (Q103789) всё равно придётся добавлять в отношении периода более раннего, чем 2005 год, конечно, проставляя квалификаторы. Страшного в таком удалении возможно и нет, но лишнюю работу делать никому не хочется. P.S. Спасибо, что быстро и понятно ответили. --Ksc~ruwiki (talk) 20:52, 19 April 2017 (UTC)

Constraint report P274 (chemical formula): error code 'badmd5'[edit]

Hi. It seems the constraint report for chemical formula (P274) has been failing since September with the message "Edit failed, error code 'badmd5' (The supplied MD5 hash was incorrect.)." Prior to this, it was untouched for a long time -- the last successful KrBot report appears to be April 2015. Do you know what causes this? Perhaps some differences or weirdness in Unicode handling. Seems like something that needs to be fixed in User:KrBot. Intgr (talk) 15:17, 24 April 2017 (UTC)

Hello, the issue is caused by symbol with UTF-8 code 0xCE 0x87. MD5 checksum is b80d5a5d9193d69ce0b1009c31587da5, but it is rejected by https://www.wikidata.org/w/api.php?action=edit&bot=1&assert=bot&format=json&utf8=true&md5=b80d5a5d9193d69ce0b1009c31587da5&notminor&nocreate&basetimestamp=2017-04-24T17:37:01Z&starttimestamp=2017-04-24T17:37:01Z&title=Wikidata:Database%20reports%2FConstraint%20violations%2FP274 request. I will continue the investigation. — Ivan A. Krestinin (talk) 17:41, 24 April 2017 (UTC)
Posed to Phabricator. — Ivan A. Krestinin (talk) 18:51, 24 April 2017 (UTC)

Source code for redirect fixing script[edit]

Hello Ivan, do you have the source code for the script that fixes links to redirects on Wikidata statements? I merge items a lot and this creates a lot of redirects, and I would like to be able to fix them quickly. Thank you, Harej (talk) 03:40, 26 April 2017 (UTC)

  • Hello, the source is here. But it is written on C++ and requires a number of custom libraries, makefiles and etc. I am not ready to publish whole bunch of sources. — Ivan A. Krestinin (talk) 20:41, 26 April 2017 (UTC)

Summary[edit]

[27]: Removed claim: Commons category (P373): Burston School Strike, несуществующий файл / missing file. Стоит пофиксить :) --XXN, 12:20, 29 April 2017 (UTC)


  • P.S. Вас могло бы заинтересовать этот таск. Даже если Infovarius работает/работал над этим, тут пахнет перманентным заданием. --XXN, 12:27, 29 April 2017 (UTC)
  • Описание поправил, а тот таск мне не по душе :) Не нравится мне самое главное искать глазами где-то в конце метки, зря что-ли во всех энциклопедиях фамилию первой пишут. — Ivan A. Krestinin (talk) 13:01, 29 April 2017 (UTC)

same object[edit]

Please, can you verify Q25710438 and Q25188681. It's the same

2A01:CB14:1E0:9900:9930:B901:4A5:2467 09:04, 5 May 2017 (UTC)

Merged. — Ivan A. Krestinin (talk) 20:25, 5 May 2017 (UTC)

Координаты из бевики[edit]

Добрый день! Не могли бы Вы вытянуть в ВД координаты из шаблона Славутасць (Храм, Помнік)? --Чаховіч Уладзіслаў (talk) 20:10, 5 May 2017 (UTC)

Please see Project Chat for spacecraft ontology discussion[edit]

Please see my comment here on your massive reverting. Thanks [28] - Fuzheado (talk) 01:44, 12 May 2017 (UTC)

Rodovid[edit]

Hello !
I am a member of the genealogical wiki Rodovid (FR) and I have seen that on 5 January 2015 KrBot added numerous links to records of our site.
If you are interested, there are other links that could be established towards pages of Rodovid, like the one for Ivan the Terrible.
Rodovid has a complete list of the records (some 30 000) that include a link to a Wikipedia page about the same person.
In Russian, you can find it starting here.
Some of these links are internal redirects to pages of Rodovid, but most are leading to Wikipedia and are beginning with "wikipedia:", "wikipedia:en:", "wikipedia:fr:", "wikipedia:ru:", "wikipedia:uk:" etc.
Normally, one person is identified by a unique number. Different records can be written about one person in several linguistic localizations of Rodovid but they will have the same number (a bit like Wikidata).
(In some temporary cases, duplicate records may have been written about the same person under different numbers.) Penkroff (talk) 22:02, 12 May 2017 (UTC)

Hello, Rodovid uses CC 2.5, WD uses CC0. Are these license compatible and allows data import? — Ivan A. Krestinin (talk) 06:58, 13 May 2017 (UTC)
I asked the question to Y.Boychuk (Baya), developer of Rodovid. Penkroff (talk) 18:00, 13 May 2017 (UTC)
My personal interpretation : CC 2.5 deals about the content, but just adding addresses with numbers probably does not involve any copyright; and the license says it does not prevent "fair use". Penkroff (talk) 23:59, 14 May 2017 (UTC)
Question from Baya on the Rodovid talk page : if these links were possible in 2015, why should the situation be different now ? 23:59, 14 May 2017 (UTC)
Previously bot imported data from Wikipedia. Now you suggest to import data from Rodovid site. — Ivan A. Krestinin (talk) 18:28, 15 May 2017 (UTC)
This list can not be used directly. It has many unreliable records. My bot need account for getting page source. Could you create it? — Ivan A. Krestinin (talk) 18:54, 15 May 2017 (UTC)
Of course, if Baya agrees. (Another procedure for membership is indicated on the Main Page of Rodovid EN.) Penkroff (talk) 01:00, 16 May 2017 (UTC)

Baya likes this idea of cooperation. Just send a message to : en.rodovid@gmail.com (indicating the name you have chosen for your account). Penkroff (talk) 18:00, 17 May 2017 (UTC)

Ok, I sent the email. — Ivan A. Krestinin (talk) 18:37, 17 May 2017 (UTC)
Done (see mail of answer). Penkroff (talk) 22:01, 17 May 2017 (UTC)

Rodovid uses very old Mediawiki version. I can not find a way to retrieve list like [29] via api.php. Could you help with this task? — Ivan A. Krestinin (talk) 07:53, 22 May 2017 (UTC)

Добавление описаний в элементах категорий[edit]

Здравствуйте. Не могли бы вы слегка изменить код бота что-бы он при добавлении множественных описаний в элементах категорий сделал бы это всё в одну правку (что-бы не перегрузить излишне историю элемента и списки наблюдения участников)? См. также обсуждение. --XXN, 16:14, 16 May 2017 (UTC)

Приветствую, давно бы это сделал, но несколько проблематично выполнить это пожелание из-за некоторых сложностей архитектурного характера. Вообще в плане категорий мой бот лишь подчищает то, что не сделали другие боты. Я не прочь, если другие боты будут более аккуратными и тогда моему просто не останется работы. — Ivan A. Krestinin (talk) 18:28, 16 May 2017 (UTC)

Special:Diff/487898337[edit]

Hey Ivan, can you please look up how your bot decided to remove the URL in this diff? I don’t see anything wrong with it… Regards, MisterSynergy (talk) 15:52, 21 May 2017 (UTC)

Read the edit summary till end and then take a look at Property talk:P973. --Edgars2007 (talk) 17:20, 21 May 2017 (UTC)
Oh I see. That wasn’t self-explaining Face-wink.svg … Thanks! —MisterSynergy (talk) 17:36, 21 May 2017 (UTC)

Q4476629[edit]

Сколько можно? ... ОКАТО и ОКТМО это про Урал (Буздякский район) Каранский сельсовет (Буздякский район)). А этой деревни уже нет...--User№101 (talk) 19:31, 23 May 2017 (UTC)

Поправил, бот не будет пытаться подобрать коды, если указано свойство dissolved, abolished or demolished (P576) или в графе "население" значится "0". — Ivan A. Krestinin (talk) 21:12, 23 May 2017 (UTC)

Constraint reports[edit]

In the constraint reports for P345 (IMDb ID (P345)) I realized that there are values that shouldn't be reported for the conflicts with report. For example, in the second section of the conflicts with constraint there is Spierig brothers (Q7577128), whose value for the IMDb ID (P345) is no value. This is usually done to prevent some bots from importing the value incorrectly and to indicate that the element doesn't had value for that property. I don't know if you agree or not with this usage of the no value option, but regardless of that the "no value" claim shouldn't be listed as a conflicting value. It's already ignored by the constraint report special page for that entity and it would be nice if it could be ignored by the constraint reports generated by your bot. Could you fix it for future reports? -- Agabi10 (talk) 19:34, 25 May 2017 (UTC)

I do not think that it is good idea to add novalue as workaround for bugs in bot`s code. This will generate unexpected number of =novalue claims. The claims will be similar to garbage. Somebody will add P345=novalue to every group of humans (Q16334295) item, somebody will add it to every taxon (Q16521)... I think bots must be more intellectual. It is very simple to add check constraint like "Conflicts with" to bot`s code. I have one more idea. @Lucas Werkmeister (WMDE): Could we integrate wbcheckconstraints to wbcreateclaim? Now the call supports checks "assert=bot", "assertuser=KrBot". Additional parameter like "assertconstraints=Q21502838,Q21510859" or "assertconstraints=all" will improve quality of the project data. — Ivan A. Krestinin (talk) 20:09, 25 May 2017 (UTC)
Regardless if it's a good workaround or not the entities with novalue shouldn't appear in the constraint report as conflicting values. If we want to prevent a property with novalue set I rather add an additional constraint to generate a report for the claims with novalue set for a given property. Anyway this is only for conflicting constraints, obviously novalue shouldn't be ignored for example in the single value constraints (I've seen today an entity with novalue and other two values for the same property). About the assertion of the constraints I don't really understand what type of checks it would be doing... The first two assertions are quite straightforward, but the last one should it be checking a given constraint type, should only check for the conflict constraint types or should work for any constraint? If the creation of the claim would for example trigger a constraint for a check if the instance is for a given type and the instance is not specified should that claim creation be prevented or not? -- Agabi10 (talk) 20:46, 25 May 2017 (UTC)

LC Auth ID[edit]

According to the property descriptions, this edit is incorrect.

The description of LCAuth ID (P244) says: "Library of Congress ID for authority control for persons, organizations and subject headings (for books use P1144)".

Before your edit, Cape Cod (Q27979486) was using LCOC LCCN (bibliographic) (P1144) because it is an instance of a book. You changed it to LCAuth ID (P244) which is for persons, organizations and subject headings. --EncycloPetey (talk) 23:38, 25 May 2017 (UTC)

Bot did such edit, because (according to description at property talk page) it was using inccorect format (Bibliographic records do not have prefixes n, nb, nr, ns, sh, sj, sn). So it automatically moved to LCAuth ID (P244). It was wrong anyway :) --Edgars2007 (talk) 06:42, 26 May 2017 (UTC)
But the change is also wrong. This property should not be used for books. --EncycloPetey (talk) 14:50, 26 May 2017 (UTC)
Ok, lets fix the issue. My bot changed ~60 items. But LCAuth ID (P244) was widely used for books before my bot edits, see [30]. Also format constraint on Property talk:P1144 does not allow codes like n85201132. Lets fix and improve constraints as the first step. Could you fix format constraint, add type constraint and conflict with constraint? I do not understand difference between these properties enough. — Ivan A. Krestinin (talk) 16:27, 26 May 2017 (UTC)
No one understands the difference in these properties! If you look at the talk pages, people keep asking question because they are confused, but no one gives answers. We may need to have a community-wide discussion to get answers. --EncycloPetey (talk) 17:35, 26 May 2017 (UTC)