Stream: containers

Topic: hacking on metadata blocks


view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:28):

https://github.com/gdcc/mdbtool is the place where the magic happens. Moving things from https://github.com/poikilotherm/dataverse/tree/7662-solrconfig into that repo so it is easier to reuse. Also after talking to Jan Range yesterday, in the middle of reshaping this from TSV only into a model based approach, hopefully making it reusable for other source and target formats

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 21:33):

Sounds good. In case you haven't seen it, a few hours ago beep posted some stuff about JSON Schema and CEDAR and TSVs, but I'm not sure if it will help your effort or not: https://github.com/IQSS/dataverse/issues/7173#issuecomment-1462301582

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:36):

Yeah, seen that. It's somewhat related, but not very strongly to mdbtool right now.

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:37):

mdbtool for instance is at least for the time being not at all related to #7173

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 21:39):

Fair enough. Thanks. Go, go go!

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:54):

Would you say 40 chars for a metadata field title is "relatively brief"? (https://guides.dataverse.org/en/latest/admin/metadatacustomization.html#datasetfield-field-properties)

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:56):

BTW it's crazy how these metadata fields are only hold together by duct tape in the main Dataverse code. There is at base no validation of a TSV piped into the API.

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 21:56):

coverage.Spectral.CentralWavelength is 35

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:57):

Is that a title? Looks like a name

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 21:57):

Oh, you're right, it's a name.

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 21:57):

(Violating the spec BTW as it contains a dot. Solr does not prohibit dots but strongly encourages you to not go beyond alphanumeric + underscore)

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 22:00):

Topic Classification Controlled Vocabulary Name is 47.

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 22:00):

curl -s https://demo.dataverse.org/api/metadatablocks/citation | jq '.data.fields[].displayName' -r

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 22:03):

Perfect. I was already starting to grep and cut the TSVs :-D

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 22:03):

OK I'll make it 60 then

view this post on Zulip Oliver Bertuch (Mar 09 2023 at 22:03):

"Brief"... :melting_face:

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 09 2023 at 22:04):

Like the Journal of Brief Ideas. :happy:

view this post on Zulip Philip Durbin ๐Ÿš€ (Apr 18 2023 at 13:47):

@Sherry Lake started a thread about hacking on metadata blocks and I just suggested that she could probably do this in Docker.

view this post on Zulip Oliver Bertuch (Apr 18 2023 at 13:54):

That's the idea! These containers are perfect for ephemeral throwaways to test these things. Break it, dump it, start over

view this post on Zulip Notification Bot (May 04 2023 at 07:05):

17 messages were moved here from #containers > mdbtool by Oliver Bertuch.

view this post on Zulip Philip Durbin ๐Ÿš€ (Jun 13 2023 at 20:52):

Is this related? Add use cases to get the tsvs of the Metadata Blocks #63 - https://github.com/IQSS/dataverse-client-javascript/issues/63

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 07 2023 at 11:13):

Huh, where are the 17 moved message?

view this post on Zulip Philip Durbin ๐Ÿš€ (Sep 20 2023 at 15:34):

@Vera Clemens just posted some code that seems related: https://github.com/IQSS/dataverse/issues/9463#issuecomment-1727612855


Last updated: Oct 30 2025 at 05:14 UTC