Suggestion: Permalinks in metadata exporters

[vera]

Overview of the Suggestion/What inspired this idea?

Looking at the metadata export for datasets with permalink PIDs, I found that the treatment of permalinks is currently inconsistent. Some export formats use the permalink itself as the dataset identifier + within the citation string, others use a link of the format $DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK.

Here are relevant snippets from each export format:

DataCite:

    <identifier identifierType="DOI">$PERMALINK</identifier>

(wrong identifierType is already being discussed here: #10759)

OpenAire:

<identifier>$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK</identifier>

Schema.org:

"@id":"$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK",
"identifier":"$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK"

DDI:

<IDNo agency="perma">perma:$PERMALINK</IDNo>
...
<biblCit>$AUTHORS, $YEARS, "$TITLE", $DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK, Root, V1</biblCit>
...
<holdings URI="$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK"/>

DDI HTML codebook: same as DDI

Dublin Core:

<dcterms:identifier>$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK</dcterms:identifier>

JSON:

"persistentUrl":"$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK"
...
"datasetPersistentId":"perma:$PERMALINK"
...
"citation":"$AUTHORS, $YEAR, \"$TITLE\", $DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK, Root, V1"

OAI_ORE:

"@id": "$DATAVERSE_HOST/api/datasets/export?exporter=OAI_ORE&persistentId=$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK",
...
"ore:describes": {
  "@id": "http://localhost:8080/citation?persistentId=perma:https://clinicaltrials.gov/study/NCT00080262",
  ...
}

What existing behavior do you want changed?

We would like to suggest that

1. the treatment of Permalink PIDs in all exporters should be consistent
2. preferably the persistent identifier of the dataset should be simply the Permalink, as in the Datacite export or in the JSON export

Any brand new behavior do you want to add to Dataverse?

none

Any open or closed issues related to this suggestion?

maybe #10615 (refactoring of Datacite export code), not aware of any others

Are you thinking about creating a pull request for this issue?

yes, we would be interested in creating a PR :)

cc @johannes-darms

[vera]

This suggestion could also apply to the citation exports:

BibTeX:

@data{$PERMALINK_$YEAR,
author = {$AUTHORS},
publisher = {Root},
title = {{$TITLE}},
year = {$YEAR},
version = {V1},
doi = {$PERMALINK},
url = {$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK}
}

EndNote XML:

<urls><related-urls><url>$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK</url></related-urls></urls><electronic-resource-num>perma/$PERMALINK</electronic-resource-num>

(By the way, two additional issues in these two exports. I've opened an issue for them here: #10769)

RIS:

Provider: Root
Content: text/plain; charset="utf-8"
TY  - DATA
T1  - $TITLE
AU  - $AUTHORS
DO  - perma:$PERMALINK
ET  - V1
PY  - 2024
SE  - 2024-08-08 14:38:10.884
UR  - $DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK
PB  - Root
ER  - 

[pdurbin]

As of this writing there are three more exporters at https://github.com/gdcc/dataverse-exporters you might want to check:

• Croissant
• DDI-PDF
• RO-Crate

[vera]

Yes, good idea. I couldn't get the DDI PDF exporter to work (maybe it's user error, not sure. I opened an issue: gdcc/exporter-ddipdf#3), but here's the other two:

Croissant:

"url": "$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK",
...,
"citeAs": "@data{NCT00080262_2024,author = {$AUTHORS},publisher = {Root},title = {$TITLE},year = {2024},url = {$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK}}"

RO-Crate:

{
  "@id": "./",
  "@type": "Dataset",
  "identifier": "$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK",
  ...
}

(To get this running, I had to disable the failing test, but I see you already opened an issue for that gdcc/exporter-ro-crate#1)

I also just added OAI_ORE above, which was missing as well.

[pdurbin]

@vera thanks for doing all that checking.

"url": "$DATAVERSE_HOST/citation?persistentId=perma:$PERMALINK",

Sorry, I'm a little confused. Is the URL above ok? Does it work for you? If not, because this example is from the Croissant exporter, please open an issue at https://github.com/gdcc/exporter-croissant

[vera]

Yes, I think after #10775 is merged, this URL works for us 👍 I will need to revisit this issue then