IQSS logo

IRC log for #dataverse, 2021-02-04

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
07:27 Virgile joined #dataverse
08:16 Angry joined #dataverse
08:16 Angry Why do I have charged on my bank statement from google dataverse
08:17 Angry Charges*
08:19 Angry No one's gonna say anything
12:18 donsizemore joined #dataverse
15:00 pdurbin sorry, you've got the wrong Dataverse
15:15 pdurbin poikilotherm2: did you make it to the meeting just now?
15:24 poikilotherm2 Yeah
15:24 poikilotherm2 It has been recorded, so don't worry
15:27 pdurbin 😅
15:35 poikilotherm2 Yesterday evening (OK I admin: night) I learned that the next Payara version will be including MicroProfile Config API 2.0
15:36 poikilotherm2 (in case someone is wondering why this is great, see the 2.0 spec release notes https://download.eclipse.org/microprofile/microprofile-config-2.0/microprofile-config-spec-2.0.html#release_notes_20)
15:37 poikilotherm2 Let me especially hightlight the "property expression" support
15:41 pdurbin Ah, this is similar to how we put fqdn in siteUrl.
16:02 donsizemore joined #dataverse
16:03 donsizemore32 joined #dataverse
16:16 poikilotherm2 pdurbin: yeah, but much more fancy this time! :-D
16:17 poikilotherm2 We can make the JDBC URL much more flexible this way \o/
16:18 pdurbin nice
16:18 pdurbin Gone are the days of non-upgrades under Glassfish. :)
17:39 poikilotherm2 pdurbin - Q?
18:11 pdurbin poikilotherm2: please go ahead
18:11 poikilotherm2 :-)
18:11 poikilotherm2 I have a chicken and egg problem
18:11 poikilotherm2 When starting the Solr instance on K8s and no Dataverse repo up and running, I loose the custom metadata fields in the XML files
18:12 poikilotherm2 So what I would like to see: a small lib that converts the configures TSV files into the XML and possibly other stuff
18:12 poikilotherm2 There is that code you wrote for the API call
18:13 poikilotherm2 Now I wonder if we could slice that part out of the codebase, put it into a separate lib module (maybe in another repo, as current pom might not be mixed with submodules) and reuse in Dataverse software plus CLI tool
18:13 poikilotherm2 No code duplication is my goal
18:14 poikilotherm2 WDYT
18:14 pdurbin Well, my first thought is: Does it make sense to shuffle that code around or should we finally switch to creating fields in Solr using the Solr API instead?
18:15 poikilotherm2 Meh. I would prefer a short term solution to get this going...
18:15 pdurbin ok
18:15 poikilotherm2 It already affected my installation a few times :-/
18:15 poikilotherm2 Of course I would prefer if Dataverse software would use the API :-d
18:18 pdurbin If it did use the Solr API... would your problems be solved?
18:19 poikilotherm2 Not so sure
18:19 poikilotherm2 It would depend on how this works
18:19 poikilotherm2 ah
18:19 poikilotherm2 wait
18:19 poikilotherm2 I need persistence anyway
18:19 poikilotherm2 So the config would be persisted with it
18:20 poikilotherm2 Currently, my idea goes in the direction of using the lib to convert a bunch of TSVs into the XML and either stuff that into a configmap, deploy via operator or similar.
18:21 poikilotherm2 Most likely configmap first, maybe created by operator later
18:21 poikilotherm2 When Dataverse uses the API, the config would have to be stored inside the collection, wouldn't it?
18:22 poikilotherm2 Not sure if one can reinitialise the config via API though
18:22 poikilotherm2 Like Solr starts with a standard config and the starting Dataverse installation then adds fields on the go if necessry
18:22 poikilotherm2 Haven't looked much into this, I suppose :-/
18:24 pdurbin I sort of imagined that Dataverse would hit the Solr API once for every field it needs to create. I don't know if there's a batch mode.
18:25 pdurbin I mean, if you wanna get crazy, you could go schemaless.
18:46 poikilotherm2 Na, staying sane for now
18:47 poikilotherm2 So what about slicing the code and stuffing into a lib?
18:48 pdurbin I mean, if you think it would help, I guess I'd suggest trying it. And if the approach is still looking good, let's have you join a tech hours or something.
19:00 poikilotherm2 OK I'll take a look at doing this.
19:00 poikilotherm2 It doesn't depend on doing intial action in the upstream project
19:01 poikilotherm2 I'll start this under GDCC.
19:01 pdurbin Are there other things that should be in a shared library?
19:01 poikilotherm2 About schemas?
19:01 pdurbin about anything
19:01 pdurbin stuff that https://github.com/IQSS/dataverse-client-java could use
19:02 poikilotherm2 An idea was about making the lib for TSVs a bit more generic, so we might have an option to one day use this to transform, validate etc. And it might be useful when we maybe some day go for JSON instead or in addition to TSV
19:03 poikilotherm2 Most likely the Java client would benefit from a shared data modell
19:03 poikilotherm2 DvObject could be used on both sides of the wire
19:03 pdurbin yeah
19:04 pdurbin I mean, that client code is pretty dormant.
19:04 pdurbin but I agree
21:44 pdurbin left #dataverse

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.