Time
S
Nick
Message
06:52 poikilotherm joined #dataverse
08:03 j-n-c joined #dataverse
08:40 stefankasberger joined #dataverse
11:04 pdurbin joined #dataverse
11:06 pdurbin
Yesterday afternoon at tech hours I went through my "thoughts on automated testing" post: https://groups.google.com/d/msg/dataverse-dev/ISot5k4VjZQ/t-hzPk8tAwAJ
11:07 poikilotherm
Mornin' pdurbin :-)
11:09 pdurbin
poikilotherm: morning. How's it going?
11:10 poikilotherm
Fine :-)
11:12 poikilotherm
Thank you
11:12 pdurbin
Tania added some screenshots to your issue.
11:12 poikilotherm
Yeah, I saw those
11:13 poikilotherm
I am not sure if I should add any comments
11:13 pdurbin
For others here, I'm talking about this auth issue: https://github.com/IQSS/dataverse/issues/5974
11:14 pdurbin
If you end up refactoring the description of the issue, I can let her know.
11:14 poikilotherm
I am also unsure what's her point in adding them... "For reference" is a term with a broad meaning...
11:14 pdurbin
She's trying to understand what you're up to.
11:15 poikilotherm
Should I add a drawing?
11:15 pdurbin
It's up to you. I'm answering her questions in Slack and in person.
11:16 poikilotherm
Any chance she might be with us here?
11:16 poikilotherm
Chatting might be easier than issue comments
11:16 poikilotherm
We could have another call, too
11:17 pdurbin
Yeah, another call might be nice.
11:17 pdurbin
Also, mockups of the UI would be nice.
11:17 poikilotherm
I see
11:18 pdurbin
The reason she put those QDR screenshots in there was that I gave it as an example of an installation of Dataverse that uses a non-supported external identity provider.
11:19 pdurbin
This is what I wrote: "if you go to https://data.qdr.syr.edu and click on "Log In" or "Register" you'll see that QDR has forked the Dataverse code to integrate a completely custom identity management solution. I believe that Oliver's solution will help with situations like this where an institution has an identity management system in place and wants to make use of it. That way their users don't have to
11:19 pdurbin
create yet another password to log into Dataverse."
11:19 poikilotherm
Err - but QDR is using Shib, as donsizemore mentioned the other day, right?
11:20 poikilotherm
pdurbin I don't think they use a fork
11:20 poikilotherm
Looking at the login URL , it links to the normal Shib auth
11:21 pdurbin
It was Shib initially. We can review what Jim said in here the other day when I summoned him. The details don't matter so much. QDR definitely runs a fork. :)
11:21 poikilotherm
https://data.qdr.syr.edu/Shibboleth.sso/Login?target=%2Fshib.xhtml%3FredirectPage%3D%252Fdataverse.xhtml
11:21 poikilotherm
Ah it was Jim who said that...
11:21 poikilotherm
Sry, my brain is dumping things randomly
11:22 pdurbin
"Sorry for being late to the party - QDR is using LDAP behind Shib and keeps user account details synced with Drupal." Jim (qqmyers) at http://irclog.iq.harvard.edu/dataverse/2019-06-27#i_99130
11:22 poikilotherm
Yeah
11:22 poikilotherm
Right
11:22 pdurbin
You don't think it's a good example?
11:23 pdurbin
It was the best I could come up with quickly.
11:23 poikilotherm
Indeed :-/
11:23 pdurbin
It's a fork.
11:23 pdurbin
The user accounts are in Drupal.
11:23 pdurbin
Drupal is their IdP.
11:23 pdurbin
Anyway. :)
11:23 poikilotherm
Right. But Shib is their SP, isn't it?
11:23 pdurbin
What's important is the concept, not the specifics, I think. :)
11:24 poikilotherm
Depends
11:24 pdurbin
Yes, they're using Shib.
11:24 poikilotherm
In my vision I see things like groups synced etc
11:24 pdurbin
sure
11:25 poikilotherm
This is currently very limited in Shib and not implemented in OAuth AFAIk
11:25 poikilotherm
IIRC with Shib you can join people to groups based on the Affiliation attribute, right?
11:25 poikilotherm
And use the IP groups
11:26 pdurbin
With Shib you can put people in a group based on which IdP they logged in with: http://guides.dataverse.org/en/4.15.1/installation/shibboleth.html#institution-wide-shibboleth-groups
11:27 pdurbin
"Support for arbitrary attributes beyond “Shib-Identity-Provider” such as “eduPersonScopedAffiliation”, etc. is being tracked at https://github.com/IQSS/dataverse/issues/1515 "
11:28 pdurbin
Anyway, please think about how you would rewrite "Auth Modes: Local vs. Remote vs. Both" at http://guides.dataverse.org/en/4.15.1/installation/config.html#auth-modes-local-vs-remote-vs-both once you have all your stuff merged. :)
11:28 pdurbin
afk
11:33 pdurbin
and/or how a future version of https://dataverse.org/software-features would look :)
12:13 donsizemore joined #dataverse
12:13 pdurbin
I guess what I'm trying to say is that the UI/UX/Design team is trying to come up to speed with what you're proposing. Again, I'm fine with continuing to answer questions. :)
13:32 pdurbin
donsizemore: mornin'. Dunno if you saw my note above but the dev team is pretty supportive of the plan I've laid out for improving automated testing.
13:39 poikilotherm
pdurbin: I just saw your comment https://github.com/IQSS/dataverse/issues/5962#issuecomment-512255410
13:40 poikilotherm
Maybe this is a use case for MicroProfile Config support?
13:40 donsizemore
@pdurbin um, i may have seen your note, but... i'm in?
13:40 poikilotherm
Having at least a tiny bit of config options is always better than nothing
15:43 pdurbin
yes but we need to get off Glassfish from 2014 first :(
16:06 j-n-c left #dataverse
17:29 amahdy joined #dataverse
17:31 amahdy left #dataverse
17:46 donsizemore joined #dataverse
18:27 pdurbin
donsizemore: do you want write access to that new "External Tools" spreadsheet I mentioned at https://github.com/IQSS/dataverse/issues/5028#issuecomment-512373100 ? Thanks for helping me notice that I hadn't opened it for comment.
18:34 donsizemore
@pdurbin oh, i just want read access. i shared the link with a few folks at odum so they may request as well, hope that's okay
18:39 pdurbin
donsizemore: of course! I just realized I forgot Co Ray Ray? Should he be on there?
18:55 pdurbin
Or is it a she?
19:16 donsizemore
@pdurbin part of why i passed it on to Odum =) though we're currently on hold pending re-hiring Andrey's position
19:43 poikilotherm joined #dataverse
19:48 pdurbin
donsizemore: cool, makes sense. I'm happy to add whatever tools.
19:48 pdurbin
poikilotherm: oh, the other thing I showed the other devs yesterday is https://github.com/dataversebot :)
19:48 poikilotherm
Good evening ladies :-)
19:49 poikilotherm
Its 10pm, kids are sleeping, my wife is out. Perfect time for coding
19:49 pdurbin
:)
19:50 pdurbin
My kids are in a camp 2.5 hours away but I'm driving there tonight.
19:50 poikilotherm
Dataversebot... Yeah... What did they say?
19:50 poikilotherm
Nice!
19:50 poikilotherm
You seem to be a great dad, driving over for the night
19:50 pdurbin
They were very distracted by dataversebot.
19:51 poikilotherm
Huh? Why is that?
19:51 pdurbin
I had to switch tabs back to "thoughts on automated testing" at https://groups.google.com/d/msg/dataverse-dev/ISot5k4VjZQ/t-hzPk8tAwAJ
19:51 pdurbin
I guess because of the avatar.
19:51 poikilotherm
LOL
19:51 poikilotherm
I still like it :-D
19:52 pdurbin
me too
19:52 poikilotherm
And they is a friendly bot.
19:52 pdurbin
:)
19:52 pdurbin
pronouns
19:52 poikilotherm
That was on purpose
19:52 poikilotherm
https://github.com/IQSS/dataverse-kubernetes/pull/80#issuecomment-510572540
19:53 pdurbin
:)
19:53 pdurbin
perfect time for coding what?
19:53 poikilotherm
I know, its just a machine, but who knows... When you guys get distracted, maybe its a she or he...
19:53 poikilotherm
https://en.wikipedia.org/wiki/Singular_they
19:53 poikilotherm
https://github.com/IQSS/dataverse/issues/6000
19:54 pdurbin
cool
19:54 poikilotherm
The they or #6000?
19:54 poikilotherm
(Bot?)
19:54 pdurbin
There's a somewhat related issue I could show you. Something that got pulled out of the current sprint.
19:54 poikilotherm
+h
19:54 poikilotherm
I'm all ears
19:55 pdurbin
this one: https://github.com/IQSS/dataverse/issues/6003
19:55 pdurbin
If dataset depositors choose a contributor type that isn't one of DataCite's contributorTypes, in Dataverse's DataCite and OpenAIRE metadata exports, map to DataCite's "Other" contributorType #6003
19:56 pdurbin
I guess I'm wondering if you/FZJ anticipate a similar need.
19:57 pdurbin
A need to modify the control vocabularies in the citation metadata block.
19:57 poikilotherm
Yes yes yes, but especially the thing djbrook mentions here https://github.com/IQSS/dataverse/issues/6003#issuecomment-511462695
19:57 poikilotherm
Quick example
19:57 poikilotherm
We have grants
19:58 poikilotherm
There already are fields for that
19:58 poikilotherm
Currently not using vocabs
19:58 poikilotherm
We can provide people at least some vocabs, but it should be possible to use free text, too
19:58 poikilotherm
(There might be a grant not included in the vocab)
19:59 poikilotherm
Also it might be better to tell people go and tell us about an unknown grant
19:59 pdurbin
Sure, that's probably even another level of flexibility beyond what I was thinking.
19:59 poikilotherm
Yeah, I see. Its adding vocab to a non-vocab field
19:59 pdurbin
yeah
20:00 poikilotherm
Let me brows through citation.tsv if I can catch sth else
20:00 pdurbin
The common one I can think of is that people will probably want to add or remove subjects, especially since Subject is a required field.
20:01 poikilotherm
Yes.
20:01 poikilotherm
That seems to be more or less it.
20:01 poikilotherm
At least from looking at citation.tsv
20:01 poikilotherm
Ah
20:01 poikilotherm
Maybe publicationIDType
20:02 poikilotherm
You remember the Software heritage archive?
20:02 poikilotherm
They use these SHA1 URNs
20:02 poikilotherm
One might want to provide a parser creating clickable links in the future for those
20:04 poikilotherm
Anyway, this might become usefull for other fields, too
20:05 poikilotherm
We think about adding more schemas for the different scientific communities, so it might get necessary to override or extend vocabs
20:05 pdurbin
I forget what publicationIDType is. I assume doi vs handle vs other stuff.
20:05 poikilotherm
Yeah, that kind of stuff
20:05 pdurbin
Yeah, I was thinking along the lines of "override", I guess.
20:06 pdurbin
I don't know. I'm not sure how best to solve it.
20:06 poikilotherm
Actually this pretty much reminds me of Hiera
20:06 pdurbin
But I'm thinking that installations shouldn't touch citation.tsv. And maybe they could specify and alternative subject or publicationIDType from a custom metadata block or something. This idea is only half-baked though.
20:10 poikilotherm
Shall I provide some more input about Hiera analogy?
20:11 pdurbin
I can't remember how I know Hiera. I feel like from Puppet somehow.
20:11 poikilotherm
Aye
20:11 pdurbin
But it's been a while.
20:11 poikilotherm
You might read up the concept quickly https://puppet.com/docs/puppet/6.6/hiera_intro.html#concept-5824
20:13 pdurbin
While I'm reading that, you can read this totally unrelated thing I heard about on a podcast recently: https://github.com/micronaut-projects/micronaut-security/issues/2
20:15 pdurbin
I took a quick look at that Hiera link. Interesting. Are you suggesting the approach or the library itself?
20:16 poikilotherm
Just the approach
20:16 poikilotherm
Hierarchies
20:17 pdurbin
Good, because I don't want to introduce a dependency on Ruby. :)
20:17 poikilotherm
I don't think the library would be usefull...
20:17 poikilotherm
Ah, JRuby, my dear friend
20:17 poikilotherm
Meh. No. Just the ideas behind it
20:17 pdurbin
:)
20:17 poikilotherm
We will definitly have a use case for these hierarchies
20:18 pdurbin
trees everywhere
20:18 poikilotherm
And AFAIK such things are not supported in Dryad, dSpace, Fedora or other big repo softwares
20:19 poikilotherm
I had a look at the micronaut stuff. Pretty cool. But as long as you are not planing to refactor Dataverse into a Micronaut Framework app, it will be hard to use.
20:20 poikilotherm
They implemented OIDC by themselfs, which is ok, but also means no reusability
20:20 pdurbin
I don't have a good sense of the other software. I went to Open Science Days in Berlin instead of Open Repositories in Hamburg. :)
20:20 pdurbin
I just thought you might find it interesting that they have OIDC support. That's all. :)
20:20 pdurbin
Anyway, a different topic. :)
20:25 poikilotherm
It definitly is very interesting. If we cannot find a suitable lib for OIDC we might rethink stealing... errr... forking their code :-D
20:26 pdurbin
great artists steal
20:44 pdurbin
I thought you said ScribeJava has OIDC.
20:45 poikilotherm
Partly!
20:45 poikilotherm
I kind of supports Open ID Connect _CORE_
20:45 pdurbin
right, right
20:45 poikilotherm
Which is mostly about using standard scopes and claims
20:46 poikilotherm
Things like Discovery, etc are not implemented
20:46 pdurbin
ok
20:46 pdurbin
at least it's being actively maintained
20:46 poikilotherm
Yeah, also it is a one-man-show
20:47 pdurbin
oh
20:47 poikilotherm
Me boss asked me to go here... http://www.library.ethz.ch/en/ms/Tagungen/Persistent-Identifiers-in-Research-Celebrating-10-Years-of-DOI-Desk-at-ETH-Zurich
20:47 poikilotherm
Would someone here be interested to go there?
20:48 poikilotherm
We could do another mini hackathon :-D
20:48 poikilotherm
Last time with pdurbin was fun
20:49 pdurbin
cool that Lars Nielsen will be there
20:51 poikilotherm
Britta Dreyer from DataCite will be there, too and Gabriela Mejias from ORCID
20:51 poikilotherm
FZJ is member in both orgs :-)
20:52 pdurbin
:)
20:52 pdurbin
I've never been to Switzerland. I bookmarked it, at least. Thanks. :)
20:53 pdurbin
new story on #dataverse2019: https://news.harvard.edu/gazette/story/newsplus/at-2019s-dataverse-community-meeting-an-emphasis-on-data-quality/
21:22 poikilotherm
pdurbin are you at least somewhat familiar with the code for DatasetField.java and DatasetFieldType.java?
21:33 poikilotherm
I'll call it a day...
21:34 poikilotherm
Cu
