[00:08:29] is there a channel for countervandalism cvn for wikidata? [00:08:55] I think we had #cvn-wikidata at some point, but it's probably dead by now [00:09:02] I haven't been there for a long time [00:09:18] ll check it out tnx :) [00:10:12] its alive :D [00:10:50] Is it? [00:11:02] cvnbot18 is operating [00:12:56] Oh, indeed… although it can't cope with the amount of edits it seems. And the edit summaries are poor (would need post processing) [10:29:16] * aude waves [10:29:24] Yeeeeeh [11:36:13] aude: fresh baked article placeholder. https://gerrit.wikimedia.org/r/#/c/269395/ nom nom [11:36:40] :) [11:39:16] aude: jzerebecki: am I right that Wikidata is now cutting a branch in sync with core ? [11:39:32] i.e. we push the new Wikidata together with the mediawiki stuff? [11:39:40] hashar: yes and no. [11:39:56] I am going to start cutting the branch in an hour or so [11:41:15] hashar: each mediawiki train has a branch of Wikidata that will be deployed with it. It may be a new branch we manually cut before the general branch cutting. or it may be the same as before. or it may be the same as before but with some backport. this week it is the last option. [11:41:30] hashar: we made a new wmf13 branch [11:41:45] (this time) [11:42:09] so you take whatever option you want but always cut a branch with the same name as the rest of mediawiki ? [11:42:19] hashar: the branch of Wikidata that is to be used is configured in tools/release.git/make-wmf-branch/config.json before the mediawiki branch cutting. [11:43:27] hashar: no it may still have a different name (the same as the week before), when we didn't change anything. [11:47:32] hashar: last week this didn't work, even though it worked for years? before. (master instead of the configured branch was used in the submodule checkout.) I hope that was just a fluke, lets see what happens today. [11:48:49] this special case for wikidata can only stop once our build stops being separate from the rest of core and extensions [11:49:15] jzerebecki: yeah the make-wmf-branch script was broken [11:49:31] it was not handling branches of special extensions properly but that is fixed [11:49:43] good :) [11:50:02] so I can get the make-wmf-branch/config.json change approved ? :D [11:50:09] and I will start cutting soonish [11:50:16] hashar: can you +2 it? [11:50:24] done [12:40:51] jzerebecki: around? [12:40:57] hoo also [12:41:01] yup [12:41:26] just wonder when we want to deploy our config change? [12:41:42] think they could go in swat or our own window befor ethe train [12:42:15] so we can at least monitor the new data type, and then placeholder would get enabled for real in teh train when the code is deployed [12:48:34] aude: do you think we need more time than when it is done during the swat? [12:48:44] otherwise lets go for the swat [12:50:23] swat is ok with me [12:50:29] and it is empty so far [12:50:38] (if it's just our stuff, i could do swat) [12:51:19] we only need more time if there is a problem to debug [13:02:06] aude: as it is all empty any time seems fine, ping me when you start [13:05:29] jzerebecki: ok [13:06:10] aude: bbl, food [13:07:07] k [13:15:48] it's a hoo! [13:16:03] hi aude :) [13:16:41] Looking forward to Saturday [13:21:15] \o/ [13:21:23] how long are you staying in berlin? [13:21:53] Either until Friday (19th) or Tuesday 23rd. [13:22:09] \o/ [13:22:26] Are you there the whole time? [13:22:34] you can join multichill and us for beers on monday [13:22:47] Oh, cool [13:22:49] i should be there the whole time [13:22:56] Didn't know he's in Berlin [13:23:29] * sjoerddebruin wish he was in Berlin [13:28:10] DanielK_WMDE: There? [13:28:54] aude: DanielK_WMDE: We should find some time to talk about how to manage infobox data schemes. Like a brainstorming session. [13:29:01] Would like to do that in Berlin and not remote or so [13:31:11] hoo: sounds good [13:31:41] daniel is sick (or at least was yesterday), so not sure he's around [13:31:47] but should be better by next week [13:32:09] According to his calendar he has time on Monday and the week after on Monday/ Tuesday [13:32:15] ok [16:10:21] Lydia_WMDE: ping [16:10:31] math data type has been enabled [17:08:10] DanielK_WMDE, thiemo: In https://gerrit.wikimedia.org/r/#/c/269386/ that the format_plain is used for machine readable export and not format_html which I assumed would be used for almost everything [17:10:52] physikerwelt_: format_plain is mostly for putting text into a a form field for editing [17:11:30] physikerwelt_: for machine readable output, you'd use a serializer to generate a JSON object, not a formatter [17:11:59] I'm a little bit confused about the statement "I think the RDF export is designed in a way that it currently *does* export the raw input as a string" [17:12:34] I was assuming that the html format would be exported and not the latex input for the rendering [17:12:52] the input string is not very helpful outside a wmf context [17:13:32] physikerwelt_: you would put presentational html into rdf? that's strange... [17:14:24] that's at least a standardized format [17:15:22] well, so is tex... [17:15:35] no [17:15:40] physikerwelt_: well, mathml is actually machine readable, and not purely presentational. so that would be ok [17:16:28] physikerwelt_: your extension controlls the rdf mapping, you can do this any way you want. But do it quickly, so we don't have a breaking change to the rdf mapping [17:16:55] there is evidence that even mathml without content markup mathml is good for certain retrieval task http://link.springer.com/chapter/10.1007%2F978-3-319-08434-3_15 [17:17:03] physikerwelt_: if the format you want is , then provide a mapping that does this [17:17:38] i'll leave it to the domain experts to decide whether mathaml or texvc is the better format to use in rdf literals :) [17:18:56] as domain export for wikidata... do you know how I can specify that in the math extension? [17:19:21] :) [17:19:54] sure. you need another callback in MathWikidataHook [17:20:11] is there an example? [17:20:25] physikerwelt_: the key is 'rdf-builder-factory-callback'. Look in WikibaseRepo.datatypes.php for examples [17:20:49] DanielK_WMDE: thank you that is exactly the kind of response I was hoping for [17:20:56] physikerwelt_: you can probably use a LiteralValueRdfBuilder [17:21:03] or something similar [17:21:44] physikerwelt_: you'd override getLiteralValue to return the MathML [17:22:04] is there a way to pass the information which xsd should be used to validate the input [17:22:39] or is there only a data and not metadata channel [17:29:19] DanielK_WMDE... I'm looking at the examples and it's still not clear to me how to pass the mathml string in way that rdf understands that this is valid xml [17:39:38] physikerwelt_: to rdf it will be a string literal. rdf knows nothing about xml, and doesn't care. [17:40:16] (many people thing that rdf is somehow bound to xml, because it can be serialized as xml - but the "natural" representation of rdf is Turtle, which has nothing to do with xml at all) [17:41:42] physikerwelt_: the example on https://stackoverflow.com/questions/15235012/how-to-embed-mathml-in-an-ontology-and-reference-parts-of-it also uses a string literal. The fact that it's MathML is expressed with the type uri for the literal. But in the turtle syntax given there, it'S just a string. [17:43:01] i don't think there is a way to directly specify an xsd. you specify a type uri. a client that knows that type may apply syntax checks using a well known schema. the type uri might even be an xsd schema url. rdf doesn't know or care. [17:44:42] DanielK_WMDE I read the post. But I vaguely remember to have discussed with Florian Rabe about a canoical way to map mathml to rdf and vice versa http://www.openmath.org/ontology/openmath.owl [17:45:41] physikerwelt_: a, sure, you can express the MathML DOM as rdf triples. But in this context, I wouldn't go that far. [17:46:04] especially given that this is more a phd thesis and not a standard [17:46:12] most consumers will want to work on a string, I assume. and constructing a DOM from a string is much easier with existign tools than constructing a DOM from rdf triples [17:46:21] yea, that too [17:46:36] interresting project, but i'd stick with using a string literal for now [17:48:03] I guess everybody who want's to process the mathml will figure out that it's mathml the the schema informatino can be inferred from the html spec [17:50:43] DanielK_WMDE... at the moment I get the impression that the monolingual text builder is simpler to use than the LiteralValueRdfBuilder [17:55:17] physikerwelt_: whatever works for you. [17:55:26] you can also just implement the interface directly [17:55:32] it's small enough [18:10:13] DanielK_WMDE... the interesting question is now testing... is there a ui where I can see how the RDF export looks like [18:12:14] Special:EntityData/Q2.nt or .ttl [18:12:19] physikerwelt_: ^ [18:13:20] indeed [18:13:25] thanks [18:13:38] physikerwelt_: https://phabricator.wikimedia.org/T126355 [18:14:00] Lydia_WMDE: https://phabricator.wikimedia.org/T126355 [18:14:18] DanielK_WMDE thank you I'll finish that after dinner... [18:15:25] DanielK_WMDE: thanks! [18:15:32] aude: great! [18:17:32] Lydia_WMDE: can you please announce? [18:17:47] aude: will do tomorrow morning [18:17:51] :/ [18:17:57] suppose that's ok [18:18:04] it is :) [18:18:08] ok [18:18:46] no one has created any math properties yet [18:27:10] DanielK_WMDE: As I said earlier (when you probably weren't around), I would like to set up a brainstorming session for Infobox data schemes when I'm in Berlin [18:27:32] I would like you and aude in there and maybe Lydia_WMDE or so [18:27:44] Anyway, we need to find a date [18:27:56] hoo: when are you in berlin? then i'll schedule [18:29:14] Lydia_WMDE: I'll arrive on Friday and will leave on Tuesday 23rd [18:31:04] ok looking at calendar [18:31:32] I have my stay in my calendar as well, for what it's worth [18:31:57] ah good [18:32:01] didn't check [18:33:03] DanielK_WMDE: are you gone most of next week or was that marked as vacation just in case? [18:34:04] hoo: if DanielK_WMDE is on vacation next week he and i will not have any overlap. so i suggest you meet with him and katie on monday and then you and me will sit together on tuesday or wednesday [18:34:41] That's sad :/ [18:34:54] we'll manage :) [18:34:55] but yeah, that makes sense [18:35:11] i'll wait for him to reply and then schedule [18:37:05] :) [18:38:10] benestar|cloud: hey :) you pinged? [18:40:34] Lydia_WMDE: Hi [18:47:56] d3r1ck: hi [18:49:28] Lydia_WMDE: tring to fix some bugs on Wikidata [18:49:55] just added the intro sentence concernin SpecialPage:Nearby, i think you filed that bug :) [18:50:07] oh great [18:50:13] I have fixed it and will submit a patch soon :) [18:51:11] awesome [18:54:41] Lydia_WMDE: you wanna review? [18:54:48] just submitted it :) [18:55:36] Is it needed on mobile? [18:55:59] hoo, Lydia_WMDE: My calendar for bext week is a bit misleading. I'm not at the office on monday. My kids are visiting. I plan to do a couple of days worth of home office work in the evenings. [18:56:30] Lydia_WMDE: made a patch for https://phabricator.wikimedia.org/T121395 [18:56:34] DanielK_WMDE: ok so can I schedule a meeting on thursday with you? [18:56:51] sjoerddebruin: Yeah it is needed [18:56:53] yay [18:57:21] DanielK_WMDE: say thursday at 10 or 15? [18:57:24] Hi DanielK_WMDE :) [18:57:30] sjoerddebruin: you want to review it? [18:57:44] DanielK_WMDE: or wednesday at 10:30 [18:58:56] d3r1ck: nope sorry [19:00:50] sjoerddebruin: :) [19:00:50] which thursday or wednesday? [19:00:55] Lydia_WMDE: DanielK_WMDE ^ [19:01:02] next week? [19:01:02] aude: next week [19:01:05] ok [19:01:11] sjoerddebruin: currently, i need to fix it now on mobile view but on Desktop, its fine :) [19:01:34] Lydia_WMDE: please did you discuss with Hoo? I Cced him in our mail thread [19:02:04] d3r1ck: no because as i said i am in the office officially only tomorrow again. i am on vacation today [19:02:27] * aude scratches head... trying to understand code with npath complexity of 25474176000000. [19:02:29] and i understand you are excited but we're all busy people so things sometimes take a bit ;-) [19:02:35] aude: ohnoes! [19:02:37] and how wikibase can hook into it [19:02:42] for our labels :) [19:03:17] Lydia_WMDE: i'm minding the kids, no time for meetings. [19:03:35] I'm available after 10pm... [19:03:43] will probably do archcom, etc [19:04:07] DanielK_WMDE: ok so no chance to scheudle a meeting with hoo|away next week? [19:04:23] no, sorry [19:04:39] email will work [19:04:41] week after? [19:04:47] yeah, the week after [19:04:53] sure [19:05:02] Monday 22nd? [19:05:29] Lydia_WMDE: Ohh, sorry, i forgot. Thanks for reminded [19:05:45] *reminding [19:05:55] hoo: soudns good to me [19:06:12] DanielK_WMDE: hoo: i'll schedule that then [19:06:21] hoo: do we want to sit down before that with aude? [19:06:31] because i am not there on monday [19:06:40] as you wish [19:07:04] Lydia_WMDE: Yeah, to iterate over what we aim for, yes [19:07:11] that probably makes sense [19:07:20] * aude agrees [19:07:27] hoo: ok then monday just you and daniel and aude, you and me next week [19:07:46] Lydia_WMDE: you are on vacation on the 22nd? [19:07:53] :) [19:08:05] yes- all coming fridays and mondays to get down way too many leftover vacation days from last year [19:08:14] oh noes! [19:08:19] yeah -.- [19:08:20] :D [19:10:18] Do we have an issue "how to join this IRC channel" somewhere? [19:10:24] For someone who hasn't used IRC before [19:10:34] s/issue/easy/ [19:11:09] yeah [19:11:11] sec [19:11:51] https://meta.wikimedia.org/wiki/IRC/Instructions [19:12:04] ah, awesome [19:12:38] hoo: DanielK_WMDE: aude: scheduled meetings [19:12:54] Thanks! [19:13:02] ok [19:59:11] Wish Magnus responded. :( [21:28:13] oh dear :( https://test.wikidata.org/wiki/Q232 [21:28:23] suppose someone tried to import the item [21:28:26] yikes [21:33:31] aude: looks like Daniel did that in 2013 [21:47:12] heh [21:47:13] https://test.wikipedia.org/wiki/Special:AboutTopic/Q2216 [21:47:25] some trouble to find a good item that doesn't yet have a test.wikipedia pge [21:47:27] page* [21:48:56] hoo: is there a reason i wouldn't be auto logged in on test.wikidata? [21:49:05] if i am logged in on test.wikipedia and test2? [21:49:21] (was it somethign to do with having the query subdomain?) [21:50:12] aude: Yes, we only auto login on www.wikidata.org anymore [21:50:18] an m. [21:50:35] ok [21:50:42] so not a session etc. bug [21:52:11] yeah, that is expected