Advice on Adding Custom Metadata Block

45 views
Skip to first unread message

Sherry Lake

unread,
Apr 17, 2023, 10:27:48 AM4/17/23
to Dataverse Users Community

Good morning,


We have  a group here at UVa who need a new metadata block. I’ve never added a new one, so would appreciate any documentation/tips on doing this. FYI - it's a CAD educational block (for teachers to share Makerspace and other manufactured educational materials) -https://dataverse.lib.virginia.edu/dataverse/CADLibrary 

 

I know there are instructions in the Dataverse docs, but thought the Community might have some firsthand tips and details.

 

One of my questions is, once I set up a metadata block (will be doing this on our test server), how do I tweak it? How do you remove a metadata block? Do you just upload a new one? OR Remove it from the database? Maybe it “depends”? If there are datasets using the block, I am sure it is way more involved to change/remove?

 

I am envisioning that once I get the block installed, to test it we would create a dataset and then I am sure I will have to re-do the metadata block, but then there would be an existing dataset with metadata in that block. But maybe if the dataset isn’t published, I could just delete the dataset and the metadata block would not “be in use”? And then..... re-do the .tsv file and re-load?


Oh, I know somewhere in all of this is re-indexing of Solr (after new block is uploaded, then again once the block is deleted or updated? and again???.....).


So I guess I am looking for the steps (in order) involved in creating a new metadata block.


And then steps to back track and re-do.

 

Any and all advice is welcomed!

 

Thanks.

Sherry Lake

UVA Dataverse

Julian Gautier

unread,
Apr 17, 2023, 2:43:14 PM4/17/23
to Dataverse Users Community
Hi Sherry,

Last summer Harvard Dataverse resolved a situation like the one you described, where:
  • We discussed and tested a new metadatablock as much as we could with managers of a Dataverse collection
  • The new metadatablock was added to the repository for that collection's exclusive use
  • A published dataset used fields from that metadatablock
  • The collection's managers asked for changes to the metadatablock
More details are in the GitHub issue at https://github.com/IQSS/dataverse.harvard.edu/issues/166 and there's more technical discussion in the associated pull request at https://github.com/IQSS/dataverse.harvard.edu/pull/167.

That PR describes some unexpected issues and how they were resolved. And later I opened another GitHub issue at https://github.com/IQSS/dataverse.harvard.edu/issues/181 to resolve another unexpected issue.

I hope this is helpful!

Sherry Lake

unread,
Apr 18, 2023, 9:35:49 AM4/18/23
to dataverse...@googlegroups.com
Thank you Julian,

The details in the Github are great. Yes, things one might not think about when creating a metadata block (how it is displayed).

I'll try to avoid those gotcha's and hopefully won't make new ones.

--
Sherry


--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/7658b42f-7d0c-49f4-95cf-ba071dc3a962n%40googlegroups.com.

Philip Durbin

unread,
Apr 18, 2023, 9:46:27 AM4/18/23
to dataverse...@googlegroups.com
Hi Sherry,

Just yesterday as Oliver and I were putting the finishing touches on the containerization proposal, we discussed this exact use case.

In years gone by I worked with our metadata expert at the time (Eleni) and had her spin up a Dataverse dev environment using Vagrant. She would author the metadata blocks in Google Spreadsheets and then download the tsv to load into Dataverse. (I took care of the Solr side back then.)

These days I think it's within our reach to have you and other metadata block hackers use Docker. (The Solr side probably still needs a little work.)

If you're feeling brave, lemme know and I can try to walk you through it!

Thanks,

Phil



--
Reply all
Reply to author
Forward
0 new messages