dataset version information not returned by Search API results

[tainguyenbui]

API Version
v. 4.18.1 build 267-a91d370

Description
The results returned by Dataverse when an authenticated Search is performed through the Search API are missing vital information to identify the version of the results.

This search is meant to return published and unpublished data

Current response

{
    "status": "OK",
    "data": {
        "q": "test",
        "total_count": 2,
        "start": 0,
        "spelling_alternatives": {},
        "items": [
            {
                "name": "Test",
                "type": "dataset",
                "url": "https://doi.org/10.70122/FK2/B3DD2U",
                "global_id": "doi:10.70122/FK2/B3DD2U",
                "description": "Some test",
                "published_at": "2019-11-21T10:31:07Z",
                "citationHtml": "Nguyen, Tai, 2019, \"Test\", <a href=\"https://doi.org/10.70122/FK2/B3DD2U\" target=\"_blank\">https://doi.org/10.70122/FK2/B3DD2U</a>, Demo Dataverse, V24, UNF:6:3P92OI0LZBPjC03KGIzYdg== [fileUNF]",
                "identifier_of_dataverse": "theam",
                "name_of_dataverse": "Theam Dataverse",
                "citation": "Nguyen, Tai, 2019, \"Test\", https://doi.org/10.70122/FK2/B3DD2U, Demo Dataverse, V24, UNF:6:3P92OI0LZBPjC03KGIzYdg== [fileUNF]",
                "authors": [
                    "Nguyen, Tai"
                ]
            },
            {
                "name": "test",
                "type": "dataset",
                "url": "https://doi.org/10.70122/FK2/21YOAO",
                "global_id": "doi:10.70122/FK2/21YOAO",
                "published_at": "2019-10-08T13:59:26Z",
                "citationHtml": "2019, \"test\", <a href=\"https://doi.org/10.70122/FK2/21YOAO\" target=\"_blank\">https://doi.org/10.70122/FK2/21YOAO</a>, Demo Dataverse, V1",
                "identifier_of_dataverse": "theam",
                "name_of_dataverse": "Theam Dataverse",
                "citation": "2019, \"test\", https://doi.org/10.70122/FK2/21YOAO, Demo Dataverse, V1"
            }
        ],
        "count_in_response": 2
    }
}

The problem we find is the fact that it is not possible to know what version of the dataset is each result unless we citationHtml or citation properties of each item.

Steps to reproduce

• Create a dataset or use an existing one
• Make a modification in the dataset, e.g. adding a new file, and DO NOT Publish the new version
• Perform an authenticated search, providing your API token, where the dataset just modified appears

Expected behaviour

• Version information is provided, for instance, versionState, and version numbers.

It would be great if it could also return the id of the dataset

Actual Behaviour

• No information about the dataset is provided

{
                "name": "test",
                "type": "dataset",
                "url": "https://doi.org/10.70122/FK2/21YOAO",
                "global_id": "doi:10.70122/FK2/21YOAO",
                "published_at": "2019-10-08T13:59:26Z",
                "citationHtml": "2019, \"test\", <a href=\"https://doi.org/10.70122/FK2/21YOAO\" target=\"_blank\">https://doi.org/10.70122/FK2/21YOAO</a>, Demo Dataverse, V1",
                "identifier_of_dataverse": "theam",
                "name_of_dataverse": "Theam Dataverse",
                "citation": "2019, \"test\", https://doi.org/10.70122/FK2/21YOAO, Demo Dataverse, V1"
            }

[pdurbin]

Hmm, good point. I assume you'd want "draft" or "1.1" rather than the database id from the "datasetversion" table. We could provide both.

It would be great if it could also return the id of the dataset

This is already possible if you set "show_entity_ids" to true. Please see http://guides.dataverse.org/en/4.18.1/api/search.html#parameters

[tainguyenbui]

@pdurbin thanks for the id suggestion, I definitely missed tat one.

In terms of the version, I think the versionState is displayed, as it is displayed for other endpoints such as the ones retrieving dataset information, where the possible values for versionState is DRAFT, RELEASED, etc. would be great. And then, if it is RELEASED then company it with the minor and major version, e.g. majorVersion 1, minorVersion 0

[pdurbin]

@tainguyenbui makes sense. Can you please update the title of this issue to be a bit more specific about the scope? So far all these suggestions seem easy to implement. Thanks in advance!

[tainguyenbui]

title updated, thanks a lot

[pdurbin]

We discussed this issue at tech hours on Tuesday and estimated it independently from #6300 but said it might be worth rolling these changes into a single pull request for both issues.