[18:45:36] I made SF a county of California, but it still does not appear in the query results on query.w.o - is there currently a delay? [18:52:08] does it only take preferred statements into account? [18:53:43] query results are cached – add a space somewhere to ensure you’re getting fresh results [18:54:01] you might still be seeing cached results from before you made your changes [18:54:05] I changed the whole query, added other information, but that shows up [18:54:29] so other info shows up, but sf does not [18:54:42] hm, then I’m not sure why, usually it updates quite fast [18:54:53] yes, that's what makes me wonder too [18:55:01] the other data was there within seconds [18:58:13] it was the preferred rank [18:58:13] thanks WikidataFacts [18:59:38] oh, okay, yw :) [19:00:13] (wdt: omits default rank if preferred rank statements exist, yes) [19:00:47] (as a shortcut replacement you can use p:P123/ps:P123 for all statements) [19:02:55] dennyvrandecic: should be no delay. What's the query? [19:49:59] aude: ping? [19:53:16] Jonas_WMDE: ping? [19:53:38] :P [19:53:56] jap [19:54:13] Jonas_WMDE: ah cool. I have some questions on QuerySamples patches [19:54:23] did you see them? [19:54:49] yes [19:55:08] so, is it just refactoring or it changes anything in the parsing? [19:56:08] the problem is that we don't have any test [19:56:28] yeah that's not good [19:56:37] that is why I am a little afraid to merge them [19:57:03] Jonas_WMDE: but also I'd want to rewrite example parsing so it'll be able to deal with SPARQL2 templates [19:58:07] yes, that is ok for me [19:58:09] by which I mean use rendered text, not source text, or extract properties from SPARQL (probably harder) [19:58:20] not sure yet which is better [19:58:49] but that may take time. Given than, I wonder if we should continue with refactoring and then work on top of it? [19:59:07] or delay it because the code may change [20:00:00] what do you think? [20:00:20] I think Thiemo just wanted to clean things up, because he knows regex very well [20:00:32] I see some of the patches are very obvious so it's probably ok to merge them [20:00:36] I am just happy that it works ^^ [20:01:04] others I'm not sure I get :) [20:01:19] and I also think the long term goal is to have a real api [20:01:27] ok, so I'll merge the simple ones and will try to understand more complex ones :) [20:01:46] Jonas_WMDE: btw, is somebody working on moving examples from mediawiki to Wikidata right now? [20:02:39] yes [20:03:12] the problem right now is that the templates q and p are using the wrong lua function which loads the whole json instead of just the title [20:03:24] ah cool. Because we'd probably have to use sparql2 template on wikidata more since it looks like Q and P are counted as heavy [20:03:58] Jonas_WMDE: exatcly. So SPARQL2 seems to avoid that. We could switch to using some other template instead of P/Q also [20:04:10] the problem is the i18n is a mess [20:04:29] I if we could use an item to store the sparql queries [20:04:35] Jonas_WMDE: oh, that's whole other mess... I don't even know how to get there [20:04:40] I wonder if ... [20:05:04] Jonas_WMDE: that'd be nice if we could store queries as item. But that would probably take time until we get that [20:05:23] we could as a first step just use a single item [20:05:26] at least if we want separate NS (which we probably do) [20:05:26] at least if we want separate NS (which we probably do) [20:05:47] Jonas_WMDE: what you mean? [20:06:13] a statment for every query [20:06:21] on one item [20:07:46] Jonas_WMDE: I am not sure UI would be good on that [20:08:15] you need something to display query nicely, and to list items/properties mentioned, and to have a name for it... [20:08:48] we could use the query ui [20:09:39] that's code mirror, it's heavy [20:10:26] not sure we want it wikidata gui... maybe yes. But integrating it may require some work I assume [20:11:46] code mirror? I was more thinking about the query examples dialog having edit functions [20:13:01] Jonas_WMDE: I don't think we need two separate editors, if we already have one editor and make it save [20:13:19] though I'm still not sure how save would know proper place to save... [20:18:12] SMalyshev: the query was select * { ?c wdt:P31 wd:Q13212489 } [20:18:37] but it didn't show up until I made Q62 P31 Q13212489 preferred [20:19:34] dennyvrandecic: ah, yes, that's how wdt: works. If you have preferred, only preferred are wdt:. If you don't have any preferred, then all normals are wdt: [20:19:54] dennyvrandecic: btw, does it really need preferred on P31? [20:20:00] thanks, that makes sense, I just didn't remember it at that moment :) [20:20:08] yeah, not sure about that either [20:20:09] I wouldn't expect to have preferreds on P31 at all... [20:20:24] I mean it's normal to be three things, very common [20:20:46] so instead of making one thing more preferred, you'd make all the statements normal [20:20:49] in this case [20:20:53] yeah, sounds fair [20:20:54] yes [20:21:14] done [20:21:26] let's see if there will be complaints [20:23:16] isn’t the “city” one redundant because “consolidated city-county” is a subclass of “city”? [20:24:13] (in fact I’d almost argue that the three statements could be combined into one “consolidated city-county of California” class…) [20:24:42] might be, but it's easier to deal with queries when it's not too deep an hierarchy [20:24:43] but SF is the only "consolidated city-county of California" [20:25:03] not sure if it warrants a specific class for that [20:25:06] yeah, that too, it's not much of a class if it has only one element :) [20:25:16] dennyvrandecic: hrm, fair enough :D [20:25:21] also, not saying that SF is a city sounds... dunno... :) [20:25:29] exactly [20:26:49] well, there are a lot of cities that are only indirectly instances of city: http://tinyurl.com/h6p9fcn [20:27:58] but not SF :D [20:28:01] kidding [20:28:27] it's a messy business :) [20:28:48] most of them seem to be obvious subclass of cities, like "city with piotr rights". "consolidated city-county" on the other side, doesn't sound that obvious too me... just a feeling [20:28:50] dunno, Q5 (human) is pretty much the only class I dare use without /wdt:P279* :D [20:29:10] yeah that's true :) [20:29:35] I think the city on SF doesn't hurt, but yeah, I see your point, you wouldn't want all subclasses materialized [20:29:56] that's because it's written by humans :) [20:30:10] I don’t think it hurts either [20:30:27] I used to remove redundant classes like that every now and then, but man, who has the time for that when there’s so much other stuff to do! [20:30:36] indeed :) [20:33:33] re things to do: this is a map of all cities with more than 1M people [20:33:34] http://tinyurl.com/hdtlauj [20:33:47] just by looking at it you can see how incomplete it is [20:52:52] oh, that's the right query. http://tinyurl.com/hksxpwh looks a lot better. yet I don't think China is complete [20:57:51] I kind of doubt RUssia is either - they got to have more that five million cities [20:58:25] in fact, ruwiki has 15: https://ru.wikipedia.org/wiki/%D0%93%D0%BE%D1%80%D0%BE%D0%B4%D0%B0-%D0%BC%D0%B8%D0%BB%D0%BB%D0%B8%D0%BE%D0%BD%D0%B5%D1%80%D1%8B_%D0%A0%D0%BE%D1%81%D1%81%D0%B8%D0%B8 [21:01:06] US is missing San Diego :) [21:01:52] the place of US Wiki Conference no less :) [21:08:10] the plot thickens... [21:25:50] any idea why this query displays "154 results" - if I add 120+33, it should be 153 http://tinyurl.com/zqszmcf (just curious what I am missing) [21:27:11] dennyvrandecic: 19th century BC appears twice [21:27:18] has two dates, apparently [21:27:36] thanks! How did you figure that out? just saw it? [21:27:46] yeah, I just looked through the list [21:28:22] lol, the German aliases on https://www.wikidata.org/wiki/Q186674 are a bit ridiculous dD [21:28:24] * :D [21:28:52] WikidataFacts: as are the Spanish ones [21:29:27] WikidataFacts: and the Bulgarian [21:29:27] a lot of them, apparently [21:29:31] those are redirects from the local wikipedias [21:29:37] sr, sv (sverige?) [21:29:37] yep [21:29:45] because they redirect to the article, they are assumed to be valid aliases [21:29:53] i think it is fair to delete them all and systematically [21:29:55] yep, a lot of languages [21:30:11] someone probably "slurped" redirects as aliases yes [21:30:24] that seems the most likely explanation [21:30:45] huh, es has “Broken Siglo XIX a C”… does “broken” mean something else in Spanish? (Google Translate doesn’t help) [21:31:00] not that I know off [21:32:22] and it's not even a word that could look like something existing in Spanish [21:36:07] bg has "1852 CE", "1889 CE", "1825 CE", [21:36:42] ca has "list of chiefs of state of the 1900s" [21:37:17] uh, make that BCE for bg [21:37:18] hr has "115th century bce" ... gnah [21:37:51] one way yo create a few thousand articles :) [21:39:01] Alphos: where is the 1852 BCE article on bg? I only see it linked to hi, it, sw, war? [21:39:05] https://www.wikidata.org/wiki/Q11165383 [21:39:27] dennyvrandecic it's an alias for Q186674 [21:39:38] ah, sorry, misunderstood [21:40:25] https://lt.wikipedia.org/wiki/1852_m._pr._m._e. [21:40:41] it says that 1852 BCE was a leap year starting on a Monday :D [21:40:55] a what what starting on a what now ? [21:41:15] ok, they already had years [21:41:21] a what year starting on a what ? [21:41:33] I am sure they also had weeks :) [21:41:42] sure how ? [21:42:25] oh, I just assumed. now you are making me look it up! [21:42:59] dude, how long have you been on wikimedia projects ? {{citation needed}} !!! [21:43:00] 10[1] 10https://www.wikidata.org/wiki/Template:citation_needed [21:43:17] probably depends on which culture “they” refers to :) [21:43:29] https://en.wikipedia.org/wiki/Week#Ancient_Near_East goes all the way back to Sumer! [21:44:10] to be fair, the lithuanian article says "according to the Gregorian Calendar", so yeah, the proleptic year Gregorian calendar year 1852 BCE sure started with a Monday, I totally believe that :D