[09:29:18] https://translatewiki.net/w/i.php?title=MediaWiki:Wbqc-constraintreport-explanation-part-two/en&diff=next&oldid=6415913 <3 [09:32:58] now that krbot moved the constraints, there's no link to a sparql query or even the constraint report :/ [09:33:09] nikki: I noticed that :/ [09:33:20] I was the one who installed them long time ago [09:34:21] they were in wikitext but it was my dream to create them with Lua [09:34:27] now we are closer... [09:41:35] we need a better way of editing exceptions too [09:42:44] PROBLEM - WDQS HTTP on wdqs2002 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.072 second response time [09:42:44] PROBLEM - WDQS SPARQL on wdqs2003 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.072 second response time [09:43:32] oh, KrBot moved the constraints? okay [09:43:45] I wanted to add some exceptions for a constraint, but I'm not going to sit here clicking "add qualifier" and typing "exception to constraint" over and over [09:44:08] I added one, I'll wait until someone comes up with something better before attempting to add more :P [09:44:12] exceptions should be avoided in general [09:44:24] quickstatements? [09:44:58] I don't know how to add qualifiers to existing statements [09:45:01] by the way, just fyi, we’re not using constraint statements yet because of build/deployment problems :( [09:45:27] and I would love to avoid exceptions, but the world doesn't care about how we model things :P [09:45:45] nikki: specifying the statement just like it’s a new one should work, right? quickstatements says “existing statments with an exacat match will not be added again” [09:45:52] so I assume you can use that to add qualifiers to an existing statement [09:46:36] I don't think that's how it works [09:46:47] I can try but I'll be surprised if it does [09:47:04] it does, I suppose [09:47:40] but maybe only if the statement doesn't have any qualifiers yet [09:47:48] so not sure as well [09:48:24] also I would love to be able to model the exceptions in a way that doesn't require listing everything individually, like "should have this property, unless the item has Px Qy" instead of listing all items with Px Qy as exceptions [09:49:23] PROBLEM - WDQS HTTP on wdqs2003 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.072 second response time [09:52:03] PROBLEM - WDQS SPARQL on wdqs2002 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.072 second response time [09:52:20] but that's not how it works, so I can't really avoid the exceptions without ditching a useful constraint which is usually fine :/ [09:57:11] quickstatements just added a new statement [09:57:35] damn [09:57:37] I guess it doesn't count as an exact match since I have different qualifiers [10:10:18] I see people have used the constraint properties on items too [10:10:36] Yeah, there was someone suprised in the project chat that it wasn't working. [10:10:40] I'm expecting a bump here though https://grafana.wikimedia.org/dashboard/db/wikidata-datamodel-statements?refresh=30m&panelId=5&fullscreen&orgId=1 [10:11:02] https://www.wikidata.org/wiki/Q840410#P275 I don't even understand what that's intended to mean [10:11:20] "except" [10:11:57] It's weird that we have https://www.wikidata.org/wiki/Property:P518 and not an invalid version. [10:12:23] what's up with the qualifier suggestions for that property though? [10:13:10] hm, for some reason it showed me a much longer list... [10:15:20] ok explain this one :P https://www.wikidata.org/wiki/Q243775#P135 [10:16:50] weird [10:17:50] oh, the "excluding" qualifier didn't have an "except" alias. that can't have helped [10:18:17] ahhh [11:20:33] PROBLEM - WDQS SPARQL on wdqs2002 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.072 second response time [11:36:20] PROBLEM - WDQS HTTP on wdqs2003 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.072 second response time [11:56:43] On wikisource there is an issue of a newspaper from 1836. Should it have P31 or P279 of the newspaper? Issue: Q19536397 Newspaper: Q1756598 [11:58:06] Maybe https://www.wikidata.org/wiki/Q732577? [11:58:12] And somehow connect with the newspaper... [12:00:50] Could one use https://www.wikidata.org/wiki/Property:P629 for that connection? [12:01:41] Or should that be an edition of the same item rather than one issue in a series of issues? [12:04:42] hmmm [12:05:01] A Help:sources was actually of help :) [12:05:18] We do have that, but this is just something that never happened before i think [12:05:45] We do have articles https://www.wikidata.org/wiki/Q22981130 [12:05:53] I guess this is a ¤b in this list: https://www.wikidata.org/wiki/Help:Sources/sv#Scientific.2C_newspaper_or_magazine_article [12:06:03] 4b, not ¤b [12:06:24] I mean, now I have an entire issue, not just an article [12:08:21] Right now there is around 50 issues from that year on Wikisource, but I was just about to upload the first issue from 1645 when was doing research (so there will be plenty of more issues) [14:18:08] PROBLEM - High lag on wdqs2003 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [1800.0] [14:18:39] PROBLEM - High lag on wdqs2002 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [1800.0] [14:36:48] PROBLEM - High lag on wdqs2002 is CRITICAL: CRITICAL: 100.00% of data above the critical threshold [1800.0] [16:08:53] So this week brought the dead of talk page constraints and WDQ. [16:16:01] sjoerddebruin: so you’re saying this will be a week long remembered? ;) [16:16:18] Well, the start of the first. :P [16:17:15] by the way, have the templates on the talk pages actually been removed? I still saw them on the few properties I checked [16:17:51] Well, between comments afaik [16:18:17] I still see them all on https://www.wikidata.org/wiki/Property_talk:P569 [16:18:32] even after purge [16:18:36] That's odd. [16:19:05] wtf, you’re right, they’re commented out in the wikitext [16:19:24] is there some template that shows the constraint statements on the talk page? [16:35:06] Template:Property documentation [16:35:14] Lucas_WMDE: ^ [16:36:09] oh, it’s part of that template? neat [16:36:20] thanks matej_suchanek [16:37:00] oh, via a Lua module [16:37:06] fascinating [16:37:39] I'm just inside that module... [16:44:18] and finally... [16:44:33] the first SPARQL report is restored [16:46:39] Interesting. [17:49:46] A small drop?!? https://grafana.wikimedia.org/dashboard/db/wikidata-dispatch?refresh=1m&orgId=1&from=now-7d&to=now [17:52:23] shocking :o [18:57:08] SMalyshev: I just saw https://phabricator.wikimedia.org/T170599 – do you mean there are statements with invalid item IDs in the database? [18:57:56] Lucas_WMDE: I don't know but looks like it... there definitely are ones with invalid units, globes, etc. but not sure about the rest. Generally constraint checker should be robust against such things I think [18:57:57] (as far as I’m aware, nothing in the extension currently looks at units, so I think at least the task you linked isn’t related) [18:58:03] that's the point of the checker :) [18:58:12] well [18:58:28] somehow it does throw there, so that's my guess at what's going on [18:58:33] I would say that’s validation, and therefore Wikibase’s job, before it ever reaches the checker :) [18:58:47] somewhere some check is missing and some bot put wrong data somewhere... maybe I'm wrong :) [18:59:16] Lucas_WMDE: yeah but what you do if Wikibase failed? right now we get exception and UBN ticket :) [18:59:38] so you are probably right that it should be Wikibase job but looks like it's not what is happening... [18:59:39] I’m looking into it… it just never occurred to me that I might have to catch exceptions like that [19:00:14] uhm [19:00:23] okay, false alarm… that’s not data from the database [19:00:34] Lucas_WMDE: check out https://phabricator.wikimedia.org/T170597 also [19:01:02] Lucas_WMDE: so where is it from? [19:01:08] the constraint parameters [19:01:14] which are currently imported from templates [19:01:26] that definitely needs try-catches… [19:01:27] sorry [19:01:35] (I’m not quite sure yet, still looking) [19:02:01] ohh ok makes sense... so if you have bad value in template... [19:02:29] ok, off to lunch, bb in 30 mins or so [20:39:47] PROBLEM - High lag on wdqs1002 is CRITICAL: CRITICAL: 30.00% of data above the critical threshold [1800.0] [21:08:25] RECOVERY - High lag on wdqs1002 is OK: OK: Less than 30.00% above the threshold [600.0] [21:20:34] RECOVERY - High lag on wdqs2003 is OK: OK: Less than 30.00% above the threshold [600.0] [21:21:04] RECOVERY - WDQS SPARQL on wdqs2002 is OK: HTTP OK: HTTP/1.1 200 OK - 13048 bytes in 0.073 second response time [21:21:05] RECOVERY - WDQS SPARQL on wdqs2003 is OK: HTTP OK: HTTP/1.1 200 OK - 13048 bytes in 0.073 second response time [21:21:15] RECOVERY - WDQS HTTP on wdqs2002 is OK: HTTP OK: HTTP/1.1 200 OK - 13048 bytes in 0.073 second response time [21:21:35] RECOVERY - WDQS HTTP on wdqs2003 is OK: HTTP OK: HTTP/1.1 200 OK - 13048 bytes in 0.073 second response time [21:35:35] RECOVERY - High lag on wdqs2002 is OK: OK: Less than 30.00% above the threshold [600.0] [22:25:36] Hello, all. Long time. :-). I would like to actually ask, what do you think of adding publicly traded company labels (e.g., AAPL, GOOG, etc.) as terms, or at least labels for terms, in WikiData? [22:30:49] Btw., I still have a confusion about WikiData, namely -- someone mentioned that concept instances for example “Mona Lisa (Q12418)” can be part of WikiData, however it is obvious, that WikiData guidelines are to accept only items of prominence, e.g. you are not supposed to register your car on WikiData. [22:31:16] I was looking for a registry for actual concept instances, e.g., where people can register actual things. What are your ideas about it? [22:34:16] (Well, not confusion, but conceptually, it seems, that WikiData is about concept-level things, rather than instance-level things. E.g., we don't register packs of milk, we register concept of milk.) [22:35:02] Can we say that WikiData is about asset classes? After all, "asset" financially, is anything that has an identity, essentially, equivalent to "item". Could say "item classes" (concepts). [22:38:39] If so, could we include prominent "item classes", e.g., "IBM Common Stock", "Alphabet Inc Class A", etc.? [22:38:57] Looks like at least for computers, e.g., we have https://www.wikidata.org/wiki/Q21414446 :) [22:42:43] Inyuki: I'm not a stock exchange specialist but it sounds like you are looking like this: https://www.wikidata.org/wiki/Property:P249 [22:43:18] So, and we have many stock markets, with different company IDs, wouldn't we integrate them like we integrate other DBs with their IDs? [22:43:51] as far as I can tell this property lets you do just that [22:44:27] and it is being used quite a lot (even https://twitter.com/Trump2Cash uses it!) [22:45:19] So, this relation is good to connect two concepts, e.g., company name, and its ticker in a stock market, but the ticker itself has to be created in WikiData then. [22:45:27] (to create binary relation) [22:46:48] Inyuki: no it does not expect an item as target [22:47:04] it expects a string (see https://www.wikidata.org/wiki/Q2283#P414 for an example of how it is being used) [22:47:24] But stocks are traded, and they are also items, not simple properties, I suppose. [22:47:34] So, item classes. [22:48:57] I don't understand what you want… I thought you wanted to "integrate them like we integrate other DBs with their IDs" [22:49:03] Different from the company itself, if the company is in multiple stock exchanges, I suppose. [22:49:36] (it's also not my domain of expertise) [22:50:24] if you want to "integrate them like we integrate other DBs with their IDs", that is precisely what is being done :) [22:50:47] (modulo the fact that this property is used as a qualifier and not a statement by itself) [22:51:23] I mean, like we do with using other databse sources, e.g., using as identifiers. [22:51:25] https://www.wikidata.org/wiki/Q729#identifiers [22:52:30] E.g., concept "animal" has Freebase ID "/m/0jbk". [22:52:41] yeah intuitively I would also have expected the ticker symbol to be the main statement and the exchange to be the qualifier, not the other way around as it currently is [22:52:57] Why can't we say "Microsoft" has Wallstreet ID "MSFT"? [22:53:40] Cause, e.g., Wallstreet could be treated as just another database. [22:53:52] Inyuki: you can! just add a statement with P414 -> Wallstreet [22:54:02] :D [22:54:07] and add P249 "MSFT" as qualifier [22:54:14] why is that funny? [22:54:38] Well, because WikiData is so general. [23:13:26] So, if a company that has stock, is a quantifiable entity, which is measured in number of stocks, how do we define the amount of it? E.g., we want to define a goal, with respect to asset, we call "MSFT", e.g., "to have the quantity of 2 of MSFT". I wonder, what would that mean to have two Microsofts? For example, we can easily understand what it would mean to have two apples (2 items of Q89). [23:14:50] From measure standpoint, it does not make sense to equate "Microsoft Corporation" with "MSFT", because, with respect to real numbers, they have different properties, so they are different concepts. [23:16:00] There seems to be a lo of dupes for Wikimedia Categores from ptwiki....any one avalible to help make a uery for it [23:17:30] suggestion items which are p31:categories; with sitelinks to ptwiki and no other wikipedia sitelinks, I can try and look it through some... [23:17:38] Can we have a Q (id) for "Microsoft common stock"? (and a new P (id), that would allow binary relation, describing "is_a_stock_of") ? [23:20:53] Josve05a: http://tinyurl.com/y9gnv9ta ? [23:22:32] with labels: http://tinyurl.com/ybtysmnd [23:22:58] ty [23:26:06] WikidataFacts: Can you add to display iems with only pt and commonswiki sitelinkslinks? [23:26:21] (I know nothing on how to work with SPARQL [23:30:37] Josve05a: sure :) http://tinyurl.com/ybjxs3ke [23:30:54] thanks! :D [23:33:15] times out :/ [23:33:40] really? it worked for me [23:33:48] but unfortunately I already closed the tab so I can’t send you the results :( [23:33:56] trying again now... [23:34:01] me too [23:34:03] hope for the best [23:34:42] Ah! It worked :D [23:35:06] ah ha, for me as well! [23:35:19] do your results start with Agudo and Aguaí, or shall I send mine to you so you have more? ;) [23:35:33] (you can also try tweaking the LIMIT in line 17) [23:35:59] starts with Categoria:Agudo and ends with Category:Brazil–North Korea relations [23:36:07] same [23:36:41] Hello [23:36:51] Hello from Spain [23:37:08] G'day from Sweden [23:38:45] I have a problem with the change of a title on a page in Wikidata [23:40:51] It's "Yesterday (Marianne Faithfull song)" that I would redirect to "Yesterday (Marianne Faithfull version)". It is posible. Sorry for my bad english. [23:43:12] actually, items on WIkidata should not have ( ... ) in the title, unless it is in the actual title of the thing. disambiguation should be done in the description [23:43:30] (edited) [23:44:24] Ough, I doesn't know that :( [23:46:45] But the particular problem on es.Wiki is that one page had two infoboxes. For example: one for the Beatles "Yesterday", and one for the Marianne Faithfull's version of "yesterday".