Expose authors from citation in API

[MortenHofft]

The dataset pages currently have a highlighted list of people in the top. This list of people is generated in the portal UI project.

Example https://www.gbif.org/dataset/75956ee6-1a2b-4fa3-b3e8-ccda64ce6c2d
shows Dutrève Bruno * Robert Solène * Thomas Milon * Anaïs Just

And then we have the citation string in the bottom which reads

citation: {
  text: Bruno D, Solène R (2016). INPN - Données flore des CBN agrégées par la FCBN. Version 1.2. UMS PatriNat (OFB-CNRS-MNHN), Paris. Occurrence dataset https://doi.org/10.15468/omae84 accessed via GBIF.org on 2021-04-27.
}

This part comes from the API

We want them to be the same
The ones in the top is wrong and the citation string is correct.
See discussion around this gbif/portal-feedback#3398

Could we expose the list of authors in the order they should appear in the citation object? Something like:

citation: {
  // the text is already there
  text: Bruno D, Solène R (2016). INPN - Données flore des CBN agrégées par la FCBN. Version 1.2. UMS PatriNat (OFB-CNRS-MNHN), Paris. Occurrence dataset https://doi.org/10.15468/omae84 accessed via GBIF.org on 2021-04-27.
  // this next part is the suggestion
  people: [
    "Bruno D", "Solène R" // possible as objects along with the contact Key from the API and their roles?
  ]
}

Or even better

citation: {
  text: Bruno D, Solène R (2016). INPN - Données flore des CBN agrégées par la FCBN. Version 1.2. UMS PatriNat (OFB-CNRS-MNHN), Paris. Occurrence dataset https://doi.org/10.15468/omae84 accessed via GBIF.org on 2021-04-27.
  people: [
    {
       abbreviatedName: "Bruno D",
       roles: ["ORIGINATOR", "METADATA_AUTHOR"],
       keys: [123], // in case we want to link to the contact
       firstName: "Dutrève", 
       lastName: "Bruno",
       userId: [] // could be  retrieved via keys if needed
    },
    ...
  ]
}

[dschigel]

Not a matter of preference, and need @ahahn-gbif here. Citation is built from the EML terms I think (same in IPT) and I guess we claim we follow the standard https://www.gbif.org/standards

[dschigel]

Looks like links from our Metadata guide don't link out to EML correctly, but Roles terms should be something to look for.

[timrobertson100]

For citations that we generate according to our algorithm the JSON approach proposed should be doable. We may need to consider the best approach when the citation is provided by the source, such as this OBIS example where the people in the citation are not registered contacts, and various abbreviations may make name matching non-trivial.

[mdoering]

As these people become the authors in the citation I would call them authors. A bibliographic citation has authors, editors and a publisher. Contributors are normally the ones which are NOT cited.

Splitting out these people is the first step into a parsed citation object. And I would suggest to design this together and base it on the same model at least close to BibTeX or CSL-JSON as desired output formats: gbif/portal-feedback#2027, gbif/portal-feedback#33

The main weakness of BibTeX is how authors are treated. They are a single string in a very odd structured syntax to be parsable, e.g. names are concatenated with a literal and. There is BibJSON to overcome this, but in the JSON world CSL-JSON has overtaken.

[fmendezh]

Dataset response has been changed to include a flag citationProvidedBySource and contactsCitation

"citation": {
        "text": "Kim J (2021). 2020_biodiversity_information_specimen. National Science Museum of Korea. Occurrence dataset https://doi.org/10.15468/7gktfs accessed via GBIF.org on 2021-05-12.",
        "citationProvidedBySource": false
      },
      "contactsCitation": [
        {
          "key": 2416932,
          "abbreviatedName": "Kim J",
          "firstName": "JunHee",
          "lastName": "Kim",
          "roles": [
            "ORIGINATOR",
            "METADATA_AUTHOR"
          ],
          "userId": [
            
          ]
        }