Time
S
Nick
Message
07:17 jri joined #dataverse
07:32 juancorr
Morning @pdurbin. File type detection works really well for us. Actually we have not many strange files. From 2017, only a GAMS (https://file.org/extension/exp) file (https://edatos.consorciomadrono.es/dataset.xhtml?persistentId=doi:10.21950/ZPRADY&version=1.1) , a ext file (https://edatos.consorciomadrono.es/file.xhtml?persistentId=doi:10.21950/AQ1CVX/UKWS0V&version=1.0) and a Rmd file (https://edatos.consorciomadrono.es/file.xhtml?persiste
07:50 poikilotherm joined #dataverse
10:03 pdurbin
juancorr: thanks!! The last link was cut off but I found it: https://edatos.consorciomadrono.es/file.xhtml?persistentId=doi:10.21950/O9TDMF/U6WSA3 . Very helpful.
10:04 poikilotherm
Morning pdurbin
10:05 poikilotherm
May I ask you some questions about custom metadata?
10:05 pdurbin
but of course
10:05 poikilotherm
:-)
10:05 poikilotherm
We will need to create our own metadata block to track datasets to funding
10:06 poikilotherm
I am aware of the citation metadata block and the fields about grants
10:06 poikilotherm
Thing is, we have a longer list of grants that should be choosable
10:07 poikilotherm
People tend to get lazy when entering this stuff
10:07 poikilotherm
And this is really bad for our reporting guys
10:07 pdurbin
Hmm, would a dataset template help at all?
10:08 pdurbin
I'm thinking that if the same group often uses the same grant it could be pre-populated. (But they could change it before hitting "save" if necessary.)
10:09 poikilotherm
Hmm this could be a solution, but there is a high danger with such free text fields
10:09 pdurbin
You want a controlled vocabulary? No free text allowed?
10:09 poikilotherm
At least we are able to provide controlled vocabulary :-)
10:10 pdurbin
:)
10:10 poikilotherm
We have https://juser.fz-juelich.de which has a curated list of funders/projects
10:10 pdurbin
50,084 grants?!
10:11 poikilotherm
Aye
10:11 poikilotherm
Lots of stuff going out there :-D
10:11 poikilotherm
+on
10:11 pdurbin
they should send you to the Dataverse Community Meeting ;)
10:11 poikilotherm
How is that?
10:12 pdurbin
since they have so much money ;)
10:12 poikilotherm
LOOOOL
10:12 pdurbin
anyway
10:12 poikilotherm
This is no money for central library ;-)
10:12 poikilotherm
We are just keeping track of things
10:12 poikilotherm
We don't need all of these 50k things...
10:13 poikilotherm
There are some very important records which are mandatory
10:13 poikilotherm
Those are called "PoF project", which is roughly translated as "program oriented funding".
10:13 pdurbin
Yeah, I guess you should consider creating a custom metadata block. That way you can add your own controlled vocabulary. The UI isn't great for 50,000 controlled vocabulary entries, though. I'm imagining a lot of scrolling.
10:14 poikilotherm
Yeah, that's why I am asking... ;-)
10:14 poikilotherm
We can shorten the list down to maybe a few dozens to hundred
10:15 pdurbin
Hmm, ok.
10:15 poikilotherm
But I haven't found docs about how to create a nice dropdown or checklist
10:15 poikilotherm
Is this UI magic with special code sauce for some metadata fields?
10:15 pdurbin
Well, it would work like any other controlled vocabulary, right?
10:15 poikilotherm
Obviously
10:16 poikilotherm
There might be interest in autocompletion
10:16 poikilotherm
But aside from that, it's just an identifier and a display name
10:17 poikilotherm
I can't tell how e.g. subjects and authorIdentifierScheme get those shiny dropdown / selection box
10:17 poikilotherm
At least there seems to be no control in the TSV for this
10:17 pdurbin
I guess it would work like "Language"? Long list there. Is that good enough? Or are you thinking you'll help us develop a better UI? :)
10:20 pdurbin
The "Life Sciences Metadata" block has some nice examples of checkboxes.
10:21 poikilotherm
Is that "biomedical.tsv"?
10:21 pdurbin
And you can check multiple boxes, if that's of interest.
10:21 pdurbin
Yeah.
10:21 pdurbin
Design Type: Case Control; Cohort Study
10:22 pdurbin
(after you hit save)
10:22 poikilotherm
Alright so when you use controlled vocabulary, those get rendered as checkboxes?
10:22 pdurbin
yeah
10:22 poikilotherm
How come that authorIdentifierScheme is rendered as a dropdown?
10:23 pdurbin
must be a different type or whatever
10:24 pdurbin
under Journal Metadata there's a drop down with "Type of Article"
10:24 pdurbin
I think the idea is that the dropdown is for when you are only allowing one selection.
10:24 pdurbin
And the checkboxes are when you want to allow multiple selections.
10:24 poikilotherm
Oh, that would make sense. Let me cross-check with the TSV
10:25 pdurbin
ok
10:27 poikilotherm
Alright, this seems to be the answer - it seams to get triggered by allowMultiple
10:29 poikilotherm
What I would really love to see: https://jqueryui.com/autocomplete/
10:29 poikilotherm
You can even let it load its data from a remote datasource
10:29 pdurbin
Can you make a pull request for it? :)
10:30 pdurbin
Or at least an issue? :)
10:30 poikilotherm
:-D
10:30 poikilotherm
I was about to ask... ;-)
10:30 pdurbin
either one is fine
10:30 poikilotherm
You are sure that there is no such issue yet?
10:31 pdurbin
I was thinking, there are two ways to get "funder" into the Schema.org JSON -LD metadata. As a curation task, you might want to create a script to copy the value from your custom metadata block into one of those fields.
10:32 poikilotherm
Hmm is there a way to override the grant fields already present?
10:32 poikilotherm
They already are of type text
10:33 pdurbin
Well, maybe you could make a pull request for that too. Sounds like some work though. I don't think you should hack on citation.tsv. It makes upgrades harder.
10:35 pdurbin
Some related issues: https://github.com/IQSS/dataverse/issues/3622
10:35 pdurbin
https://github.com/IQSS/dataverse/issues/2603
10:36 pdurbin
https://github.com/IQSS/dataverse/issues/350
10:39 pdurbin
about autocomplete, or typeahead or whatever, multiple select
10:39 pdurbin
Please feel free to create a new issue though.
10:39 poikilotherm
Alright :-)
10:40 jri joined #dataverse
10:40 pdurbin
as many as you like :)
10:40 poikilotherm
Hmm #2603 seems to be a close fit :-D
10:41 poikilotherm
Including #4772
10:45 pdurbin
RightInTwo is Jonas.
10:45 pdurbin
We never did play basketball.
10:47 pdurbin
The three of us.
10:48 poikilotherm
Most certainly I wouldn't. I am pretty bad at playing games where a ball is involved in any way.
10:48 jri_ joined #dataverse
10:49 poikilotherm
I have no answer from Giacomo yet. (And I doubt SSI would provide funding for this...)
10:49 pdurbin
Ah, so maybe you didn't appreciate my tweet. :)
10:50 pdurbin
I've still been meaning to write him back. Should I reply all on the thread?
10:50 pdurbin
the hackathon in Europe isn't going to happen without a champion
10:51 poikilotherm
Yeah :-(
10:51 poikilotherm
Maybe I could ask people for us getting involved in this when we are in production...
10:51 poikilotherm
For now, this seems to be a bit early
10:52 poikilotherm
And everybody is busy with WissKom conference ;-)
10:53 pdurbin
looking forward to having you on the map
10:53 pdurbin
and the new metrics page
10:53 poikilotherm
Oh by the way: do I have it correct that people coming from Shibboleth auth will not see a signup form as people coming from OAuth will?
10:53 pdurbin
yeah, it's a different UI/UX, unfortunately, older
10:53 pdurbin
we did shib first
10:53 poikilotherm
This is a very usefull information
10:54 pdurbin
someday we'd like to re-do shib to be more like oauth
10:54 poikilotherm
People start political discussion here about what auth strategy to use... :-/
10:54 pdurbin
especially from the technical side, to be less strict about required shib attributes
10:55 poikilotherm
That really would be a benefit :-)
10:59 pdurbin
just let the users fill in an email address if shib doesn't provide one
11:00 jri joined #dataverse
11:03 jri_ joined #dataverse
11:35 poikilotherm
Oi pdurbin, I just stumbled over you cherry-pick question. See https://github.com/IQSS/dataverse/pull/5834
12:00 poikilotherm
Oi pdurbin I think I just found a regression... Lets see if there is already an issue for this
12:08 pdurbin
poikilotherm: thanks! I made a small change to that pull request and moved it to QA.
12:08 pdurbin
What's the regression?
12:09 poikilotherm
You cannot change metadata fields for dataverses
12:09 poikilotherm
I tried in our installation and on demo.dataverse.org
12:09 poikilotherm
This is my test dataverse: https://demo.dataverse.org/dataverse/testiotis/
12:10 poikilotherm
In our installation I tried on root dataverse
12:10 poikilotherm
Looking into the fields of "citation", all are defined as "show", none "hidden". But when adding a new dataset, they don't appear
12:11 poikilotherm
When adding other schemas, at least in our installation, they do not show up in the "add dataset"
12:11 poikilotherm
Ah
12:11 poikilotherm
Maybe this is by design
12:11 poikilotherm
Lots of 'em are "displayOnCreate=false"
12:12 poikilotherm
Yeah, they show up when editing the dataset metadata
12:13 poikilotherm
Let's see if making them required lets them show up
12:13 poikilotherm
(on creation)
12:14 poikilotherm
Ok, that works.
12:14 poikilotherm
Maybe there should be an option while editing the metadata fields that they should show up on creation
12:34 pdurbin
poikilotherm: sure, good idea. Please open an issue about this.
12:35 poikilotherm
Haven't scrolled through the docs yet, if the design is documented... ;-)
12:40 pdurbin
suggestions (such as a "design" section) for a future version of this page are very welcome: http://guides.dataverse.org/en/4.14/admin/metadatacustomization.html
12:44 poikilotherm
Alright, that's it for me for today... Going to pick up a core drill and make some holes at home...
12:44 poikilotherm
Will open an issue on this on Monday... ;-)
12:45 poikilotherm
Have nice weekend
13:02 donsizemore joined #dataverse
13:10 donsizemore
@pdurbin on redesigning shib: in his wholetale issue craig would like to use cilogon
13:11 donsizemore
@pdurbin list of current identity providers at https://cilogon.org/
13:59 jri joined #dataverse
14:16 JimM joined #dataverse
14:17 donsizemore joined #dataverse
14:28 donsizemore
@pdurbin also, if you're curious, i've updated the beloved zipfile. i can send it to support (or open an issue) if you think either would be helpful. i can also be quiet if it's just a "me" problem
14:37 pdurbin
donsizemore: morning! I definitely want to get the api test suite working from https://jenkins.dataverse.org and I'm happy to help. CILogon isn't really on my radar much. Thanks for the link and for mentioning the issue Craig opened: https://github.com/IQSS/dataverse/issues/5455
16:26 donsizemore joined #dataverse
16:34 pdurbin
donsizemore: knock knock
17:50 pdurbin
Dataverse 4.14 is out: https://github.com/IQSS/dataverse/releases/tag/v4.14
18:15 donsizemore joined #dataverse
18:15 donsizemore
@pdurbin just signing in to bump =)
20:39 pdurbin
have a great weekend, everyone
21:13 jri joined #dataverse
23:29 jri joined #dataverse
