[09:10:26] https://phabricator.wikimedia.org/T104833 [09:33:33] Nemo_bis: Same for http://wdq.wmflabs.org/stats, few days lag. [11:23:56] * hoo waves [11:25:33] @HPI-Students: ping [11:26:42] hi, I am using the API to check Wikidata edits (https://www.wikidata.org/w/api.php?action=help). I have a question about the feedcontributions action: when I check my contributions in the UI I see that I created an item. However, when I check via the API (feedcontributions action) I see the s referring to the edits of the label and description (with "wbsetdescription" and "wbsetlabel" under the element), but not [11:27:06] hoo: pong [11:27:08] Thiemo_WMDE: https://phabricator.wikimedia.org/T102992 [11:27:30] tamslo: do you know if the comments in https://phabricator.wikimedia.org/T102992 were addressed [11:27:33] ? [11:27:42] tamslo: Ok... escaping has been altered again... have you dobule checked that? [11:27:53] I looked over it and it look ok, but I didn't check *all* function calls [11:27:59] * looked [11:28:12] can I get the information about a user creating an item within the list of contributions of a particular user? [11:28:51] The comments in the Ticket were addressed [11:29:35] And there shouldn't be double escaping anymore [11:29:38] tamslo: Ok [11:30:13] Also you altered the DB schema... do we need to perform the schema changes on testwikidatawiki? [11:31:23] tamslo: [11:32:21] tamslo: it doesn't look like https://phabricator.wikimedia.org/T102992#1403972 was in any way addressed, sorry that I only notice this now [11:32:35] * points people to https://gerrit.wikimedia.org/r/#/c/222926/ too* the removal of that class is in WB master now, so that should also be merged soon! [11:33:56] DB schema change is not important (I was told) [11:34:31] It looks backward compatible [11:34:40] We can just re-create the table on testwikidata [11:34:44] but not now [11:34:50] to many moving pieces [11:35:41] The ticket was addressed, in one of the last changes that were merged ... I cannot look it up, sitting with my cell phone in our presentation [11:37:03] I'm not super happy with the database stuff, but let's not block on that [11:37:15] The tables are so small, we can just re-create them at any time [11:37:26] (Not nice, though...) [11:38:01] I don't know what was done explicitly so I unfourtunately can't tell you much about it ... [11:39:27] hoo: yes lets proceed as planned [11:40:07] On taht :) [11:42:25] jzerebecki: Sanity check https://gerrit.wikimedia.org/r/223011 please [11:45:00] hoo: +1 [11:45:08] thanks :) [11:50:51] I just created the constraints csv for wikidata... got 5311 rows... is that ok? [11:50:58] Oh, they're gone :S [11:52:02] <_tamslo> 5311 sounds good, gone doesn't [11:52:16] Oh, didn't see that nick [11:52:18] ok :) [11:53:19] Sorry, IRC on cellphone is not cool ... [11:53:33] scap is running, guess it will be done by 14:15 [11:54:59] hoo: awesome :) [12:15:29] Scap is done [12:16:11] Escaping looks good now https://test.wikidata.org/wiki/Special:ConstraintReport/Q22 [12:16:45] Enabling on Wikidata now [12:17:37] Sounds very good :D [12:18:26] done [12:18:31] Will populate the table [12:18:36] Yaaay :) [12:19:40] Done [12:19:41] please test [12:20:12] https://www.wikidata.org/wiki/Special:ConstraintReport/Q4115189 [12:20:14] :) [12:20:46] Shouldn't there be tooltips? [12:20:55] modules seems to be missing (css and js) [12:21:03] uh [12:21:11] weird [12:21:54] Let's just give RL its 5 minutes [12:23:02] Works now [12:23:30] nice! [12:24:43] Error logs are empty so far, so I guess we can call this a success \o/ [12:25:04] AWESOME :D thanks a lot!!! :)) [12:26:03] Wuhuu! :( [12:26:07] * :) [12:26:20] Thank you all! :D [12:27:07] You're welcome :) [12:27:31] I'll be around for another bit, ping me should there be issues [12:27:53] hoo: you need tell them a shipping address for all the bribes ;) [12:29:49] :D [13:54:52] Have to leave, my notebook is dying [13:55:50] * hoo had a CPU core go offline :S [14:02:32] My CPU is strained too [18:33:34] hey hoo :) [18:33:50] hi benestar :) [18:33:55] Still in Berlin? [18:34:27] hoo: nope, just arrived at home :) [18:52:01] ohia benestar [18:52:49] benestar: I have a change you could make to DM/serialization if you want :P [18:53:19] hmm, actually might require a bit more lookking into... [18:56:14] sure [18:56:17] tell me :) [18:57:46] Lydia_WMDE: If you need an example of downstream usage, check out http://bengdb.beeldengeluid.nl/227799 [18:58:07] It's partly based on wikidata, raw data at http://bengdb.beeldengeluid.nl/227799.json [18:58:24] well benestar I think I am about to push a new version of the use DM serialization in RB for references [18:58:44] but there are 2 rather large TODOs / hacks in there, which probably indicate things lacking in DM serialization currently [19:02:26] addshore: sorry, I don't get it o.O [19:02:33] give me a second [19:03:02] what is RB? [19:03:07] ResultBuilder? [19:03:28] ya [19:03:42] basically look at https://gerrit.wikimedia.org/r/#/c/222622/2 and the code added in ResultBuilder [19:03:53] there are 2 rather large comment block documenting the hacks.... [19:04:26] :D [19:04:33] but now I can move on the the next bits! :D [19:15:41] addshore: reference hashes might be useful for further editing of that reference [19:15:53] yes reference hashes are usefull [19:15:55] but not snak hashes [19:15:56] I believe we require the reference hash at some api point [19:16:14] oh, it's the hash of the snaks... [19:16:15] infact having the snak hashes might confuse people (as they are not needed at all for the api) [19:16:20] yup :) [19:16:32] well, imo we should serialize snaks everywhere in the same way [19:16:43] so either include the hash everywhere or ignore it [19:16:59] otherwise, it might also be confusing that snaks are serialized in different ways [19:17:20] hmm, I think it might be simpler to have a single hash there though [19:17:36] I think there should be an option for having hashes in snak serialization or not having them [19:17:47] or we could just add the hash to the output, I dont actually care that much [19:18:50] anyway, Ill continue with the chain and come back to it at the end :D [19:22:01] sure, that detail shouldn't really block us [19:22:15] silly phpstorm doesn't allow to open several projects in one window [19:22:38] just open different windows? :P [19:23:16] gnah :S [19:24:22] ya :P [19:24:47] btw we already have serializers for TypedSnak [19:24:53] but they aren't used afais [19:25:54] Yes, I mention them in my comment [19:26:08] The issue is, I dont have typed snaks :P [19:26:21] and typed snaks can not be added to a Reference object [19:26:30] thus I wouldnt be able to just serialize the reference :/ [19:26:31] addshore: we need a service to turn snak objects into typed snaks [19:26:37] should go into WikibaseDMService [19:26:38] *s [19:26:46] Well, we would have to change how TypedSnaks works too [19:26:56] TypedSnakReference? ;) [19:26:57] why? [19:27:09] TypedSnak implements nothing [19:27:17] cant be passed into a raference and then serialized [19:27:19] just create a service that turns Snak objects into TypedSnaks [19:27:37] but I still want to call ->serialize ( Rference $reference ) [19:27:37] so make it implement snak? :P [19:27:53] thats your job ;) [19:28:01] you make that work, im going to do claims ;) [19:28:38] just got addClaim addCliams and addEntityRevision to go [19:28:38] sure [19:32:31] addshore: that is going to be super ugly [19:32:41] which bit? [19:32:46] all :P [19:32:48] I have to add the same hack in for addClaim btw [19:32:57] and will probably have to for addClaims and EntityRevison too [19:33:03] just wrapping all the methods in TypedSnak which are defined in Snak is already evil [19:33:13] mhhmhmmm [19:33:26] but then we would have to do a `instanceof TypedSnak` check in the SnakSerializer as well -.- [19:33:48] unless we do somehting properly after serialzation [19:33:57] something that is actually made to inject it into the serialization... [19:34:05] or use the slotty thing? [19:34:41] could also use the slotty thing xD [19:34:52] basically TypedSnak is already a slotted snak [19:35:00] with one fixed slot [19:35:13] yeh.. [19:35:23] except the slot is a string instead of a DataValue [19:35:37] addshore: a StringValue ;) [19:35:53] however, the slotty stuff only works for custom value snaks [19:35:56] mhhm, they are serialized different though [19:35:58] not for novalue or somevalue [19:36:15] currently they aren't serialized at all... [19:36:18] well currently the datattype is only in value snaks, not no or some value [19:36:25] how should the slots be serialized in your opinion? [19:36:35] mhhhm, no idea, just a string? :P [19:36:47] addshore: I guess that won't work [19:36:58] because they might contain more complex information [19:37:18] * benestar isn't sure about that.. [19:37:55] so yes, this won't the same way we insert datatype information currently [19:38:20] however, we could handle "datatype" as a special slot in serialization [19:38:31] with some extra handling <- super ugly [19:41:30] * benestar is still not happy that DerivedPropertyValueSnak extends PropertyValueSnak [19:53:05] addshore: you still didn't merge my patches yet =o [20:15:25] benestar: I refactored my hax https://gerrit.wikimedia.org/r/#/c/223101/2/repo/includes/api/ResultBuilder.php [20:15:40] claims are now done... just entity revisoon to tackle now [20:17:45] YAY https://usercontent.irccloud-cdn.com/file/1oQwgzeN/ [20:18:09] of course I forgot that the entity stuff uses the claim stuff... [20:23:15] infact, no, I have no idea why they are failing,... [20:26:00] ahh, edit entity, which the tests use, bah [20:43:59] benestar: so yes, the only 3 problems I have run into with only entities left is snak hashes, datatype injection and now something to do with grouping lists https://integration.wikimedia.org/ci/job/mwext-Wikibase-repo-tests-mysql-hhvm/2285/console