[05:49:09] <tanny411>	 Where can I learn about the available WDQS service? Looking for gas'
[05:49:20] <tanny411>	 'gas' now.
[05:49:27] <tanny411>	 Like this one: https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service/queries/examples#Asterophryinae_parent_taxon_reverse_graph
[06:21:47] <dcausse>	 tanny411: it's a mix of features provided by us and blazegraph, gas is provided by blazegraph it means Gather, Apply, Scatter originally invented by PowerGraph (https://blog.acolyer.org/2015/05/29/powergraph-distributed-graph-parallel-computation-on-natural-graphs/)
[06:22:33] <dcausse>	 I can list what we provide but I'm not sure we have a exhaustive list of what's available in blazegraph
[06:25:10] <dcausse>	 note that federation also uses "service" but this one might be properly parsed by jena (e.g. SERVICE <http://dbpedia.org/sparql>)
[06:32:32] * tanny411 ah, okay. thanks!
[06:33:00] <tanny411>	 dcausse: I didn't get what you said about service though.
[06:47:46] <dcausse>	 tanny411: I'm building a small sheet with all I know
[06:49:00] <tanny411>	 Thats very helpful, thanks!
[07:00:16] <dcausse>	 tanny411: I might have miss some: https://docs.google.com/spreadsheets/d/1ZVZDtkBAnl0RByQSqtFyooT1ba5X0Cm-GiKKpwpdY_o/edit#gid=0
[07:03:26] <tanny411>	 dcausse: requested for access.
[07:03:31] <dcausse>	 oops sorry
[07:05:10] <tanny411>	 Looks great!
[07:05:45] <dcausse>	 if you a list of failed queries feel free to share I can take a look and see what I missed
[07:06:01] <dcausse>	 if you *have
[07:14:25] <tanny411>	 Will do
[07:22:30] <zpapierski>	 dcausse: can I help in anyway to push streaming updater work ahead?
[07:22:58] <dcausse>	 zpapierski: I was about to investigate https://phabricator.wikimedia.org/T287969
[07:23:05] <dcausse>	 last known blocker
[07:23:29] <dcausse>	 I suspect java11 stricter policies but could be packaging issues
[07:23:42] <zpapierski>	 ok, I'm on it
[07:23:45] <dcausse>	 thanks!
[07:24:26] <zpapierski>	 yw
[07:24:40] <zpapierski>	 I'm going to push this through process, since it's a blocker
[07:24:50] <dcausse>	 sure
[07:25:14] <zpapierski>	 (we probably should reorganize our backlog and Ready For Development due to priority shift)
[09:47:31] <gehel>	 Lunch
[10:34:16] <zpapierski>	 dcausse: I see you followed up with converting all the deployment script for streaming udater to Python, nice!
[10:34:41] <dcausse>	 yes bash started to be a big mess
[10:35:39] <zpapierski>	 yeah, I wasted some time trying to figure out how to parse yaml in Bash, until the perfect solution presented itself - "just don't"
[10:38:22] <dcausse>	 we should find a repo so that we can write tests cause it's starting to be too big for a simple standalone script
[10:38:30] <dcausse>	 lunch
[11:00:23] <zpapierski>	 true
[11:00:59] <zpapierski>	 importing flink swift project took a very long time...
[12:14:28] <zpapierski>	 dcausse: I'm assuming that swift plugin doesn't need to be compiled to 11 targer?
[12:14:33] <zpapierski>	 s/targer/target?
[12:20:17] <dcausse>	 zpapierski: no clue but I doubt since there's a single flink package to download (only scala version differs)
[12:20:46] <zpapierski>	 makes sense
[12:23:36] <zpapierski>	 anyway - fix is easy in this case and you basically already suggested it - adding the offending class (or rather its copy) to our plugin
[12:23:57] <zpapierski>	 where did you test that?
[12:27:52] <dcausse>	 on k8s, I'm afraid it might be hard to test locally
[12:28:11] <dcausse>	 that requires building a new image, shipping the chart and redeploying the application
[12:28:26] <zpapierski>	 sounds fun
[12:28:40] <dcausse>	 testing this locally would require setting up swift
[12:28:42] <zpapierski>	 I'll assume for now it worked (can't imagine why it wouldn't)
[12:28:54] <zpapierski>	 and ship the new version
[12:29:17] <zpapierski>	 right, I was thinking Yarn, but JDK11 is a no-go there
[12:29:55] <dcausse>	 perhaps rename the package to trigger compilation failure for all other offending accesses
[12:29:55] <zpapierski>	 sonar cloud doesn't like me putting a data access class without adding tests for it...
[12:31:08] <zpapierski>	 I actually led with that
[12:32:03] <zpapierski>	 (that's why I think it should work)
[12:32:41] <dcausse>	 ok then let's build all this and ship
[12:33:57] <zpapierski>	 https://gerrit.wikimedia.org/r/c/wikidata/query/flink-swift-plugin/+/710018 - Sonar complains, but since I want to keep it close to the original, I want to ignore it this time
[12:34:02] <dcausse>	 btw having https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/708974 might help so that generated patches by the blubber pipeline should be correct
[12:38:48] <zpapierski>	 ah, I need to drop off for now - I'll pick it back up tomorrow (it's like 3h of meetings for the rest of my day today)
[14:56:37] <dcausse>	 hm.. we have 2 office hours calendar invites should we drop one?
[14:58:13] <dcausse>	 removed
[14:58:28] * ebernhardson waits to see which is the chosen one
[19:33:25] <gehel>	 ebernhardson, ryankemper: meeting? https://meet.google.com/yzs-ckdg-bba
[19:33:45] <ryankemper>	 oh right, brt
[20:19:19] <ebernhardson>	 doh!
[20:19:43] <ebernhardson>	 sorry :(
[20:30:00] <bd808>	 ebernhardson: my sketchy notes are at https://phabricator.wikimedia.org/T271376#7261270. If you think of something to add please do comment on the task or tell me.
[22:28:14] <mpham>	 in case yall didn't see, clara has provided a generated sample data for for the add image task that we are now able to test with: https://phabricator.wikimedia.org/T285816#7259748
[22:32:25] <ebernhardson>	 cool! in a quick look the format seems to be as requested, same shape as another one we already process
[22:34:45] <mpham>	 oh that's great! no secret surprises