[08:11:25] https://html.spec.whatwg.org/multipage/dom.html#the-dir-attribute
[08:11:27] Thiemo_WMDE:
[08:11:29] jzerebecki:
[08:31:28] moin :)
[08:55:51] hey aude
[09:01:19] o\
[12:49:47] DanielK_WMDE: https://github.com/wmde/WikibaseDataModel/pull/665
[13:07:53] are there things you can do with the wikidata query service that cannot be done with the API?
[13:20:13] heyoni: yes, the query service allows you to construct quite complex queries and get an answer quickly, for example if you wanted to find how many people were born in a specific year, the only way to do that via the api is to fetch all items linking to Q5 (human) and check each one to see if it's an instance of human (and not linking to the item for another reason) and then check whether there's a match for the date of birth. there are over 3 million
[13:20:13] people, so at 50 per api request (or whatever the limit is), it would take you ages to get the answer
[13:22:28] or if you want to do something like find all items within x km of somewhere, you'd have to fetch *all* items with coordinates (also millions) and filter out all the ones which don't match, whereas the query service has a way to search for things like that
[13:23:38] nikki: i see...
[13:23:39] I guess technically nothing is impossible via the api as long as you have unlimited time and bandwidth to fetch as many items as you need to, but it's so inefficient that you'd be crazy to do that when you could use the query service or download a data dump :)
[13:26:48] would it be a simple api request to say, search the database for all items with property x = something?
[13:27:13] or would that require getting everything with property x, then filtering those equal to what i want?
[13:28:07] i ask because i'm looking and it doesn't seem obvious (from the sandbox at least), and what you described seems like a job for the query service
[13:31:28] the query service would be better there, yes. the api can't search for things using a property, only for pages which link to other pages (which for items does mean it's using the property in some way) but you can't filter by the values
[13:33:20] I think the one place where the api is better is for text searches, like finding items with a particular word in the name, but I don't do things like that very often, so I don't have much experience with that
[13:33:51] good to know...i was doing just ignoring the query service completely...i think because freebase had a way of doing some really complex things via their api, and that's where i came from
[13:36:29] ah :) I never used their api so I'm not familiar with what it can do, but for wikidata the api is basically the mediawiki one with a few extra things added on, it was never really designed for querying structured data, whereas that's the whole point of the query service
[13:58:39] * aude waves
[14:18:42] Thiemo_WMDE: easy chocolate (once jenkins runs) https://gerrit.wikimedia.org/r/#/c/292142/
[14:30:06] * aude off ....
[14:30:14] * heyoni wave
[20:15:18] jzerebecki: tody's rfc meeting seems to be all about security https://phabricator.wikimedia.org/E198
[20:20:29] DanielK_WMDE__: I noticed. I got a mail about the change without being shown as subscribed... I'll be there.
[20:31:09] cool. not sure how closely i will follow myself. i'm pretty shot