Dataverse: Add datacite metadata in schema.org/JSON-LD format to dataset page via DataCite service

[jggautier]

Related to #2243, being a quicker, less inclusive solution to adding schema.org metadata to most dataset landing pages

[jggautier]

The script gets the DOI from the page via a "DC.identifier" meta tag

@scolapasta Now that dataset pages will have these meta tags in 4.7, could we use DataCite's service as a first step toward getting metadata in schema.org/JSON-LD format in dataset pages?

[pdurbin]

The service is now in beta, and an example with a Dataverse DOI would be https://data.test.datacite.org/application/vnd.schemaorg.ld+json/10.7910/DVN/DA1WTL

@mfenner the URL above now returns "500 Internal Server Error". Is there a working example somewhere? Thanks!

[pdurbin]

@jggautier found https://data.datacite.org/application/vnd.schemaorg.ld+json/10.7910/dvn/icfngt as an example from https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/ICFNGT . Thanks!!

@jggautier also explained that the idea may be to download the JSON above from DataCite and simply present it for download from Dataverse. I'll play around with it.

[jggautier]

Thanks Phil! (How did you know to remove the "test" from that URL?)

The idea is to take the JSON from DataCite and plug it into the header of dataset html, like the metatags with the Dublin Core metadata.

[pdurbin]

The idea is to take the JSON from DataCite and plug it into the header of dataset html

Ok, so I shouldn't make the "schema.org/JSON-LD" metadata available via the Export button like this which seems to be what #3700 is about:

What tools or services will consume the JSON in the header of the dataset HTML?

Also, how is this issue different than #2243?

[pdurbin]

What tools or services will consume the JSON in the header of the dataset HTML?

I think I found a good example:

https://search.google.com/structured-data/testing-tool#url=https%3A%2F%2Fsearch.datacite.org%2Fworks%2F10.7910%2Fdvn%2Ficfngt

In short, I think I could try to use this as a model: https://search.datacite.org/works/10.7910/dvn/icfngt

[csarven]

http://rdf.greggkellogg.net/distiller can probably check for JSON-LD in HTML. See also http://osds.openlinksw.com/

[borsna]

What tools or services will consume the JSON in the header of the dataset HTML?

@pdurbin crawlers like google, bing, yahoo etc will look for it in the header. not sure if having it as an export target is necessary.

Its not different to have it inline (suggested in #2243) or declared as json-ld in the header.

The dateset json-ld example looks like a good example 👍

[pdurbin]

@csarven thanks! https://addons.mozilla.org/en-US/firefox/addon/openlink-structured-data-sniff/ seems nice:

[pdurbin]

@jggautier I have metadata questions for you. Please see 1b62596 and the screenshot below:

[jggautier]

Thanks for stopping by Phil! From our brief discussion, it sounds like you're not using the DataCite service, which @scolapasta suggested only as an intermediate step, but writing code so that the application can transform the metadata itself, what we wanted in the end. Should we move this discussion to #2243 and close this issue?

[pdurbin]

At standup we decided we're doing #2243 rather than this issue. Closing.