Determine how citation is generated for DMPs and why it differs within tool

[bofstein]

We have had numerous people reach out to ask about how the citation for a DMP is generated, as it often isn't what people expect. The difficulty lies in the potential non-overlap between the research Project Contributors and the DMP Collaborators, and people have different opinions on who should be an author on the DMP.

Without making a determination of what the authorship should be for now, we have the following issues to gather information on:

• Some published DMPs don't list a citation at all, like this one: https://dmphub.uc3prd.cdlib.net/dmps/10.48321/D14F38aa13
• Some published DMPs have unclear citations, such as excluding the primary contact, like this one: https://doi.org/10.48321/D1B581751F or excluding the owner like this one: https://doi.org/10.48321/D1A44DD570
• DMPs have a different citation showing on the plan page (the cases I saw just showed one author on the Finalize tab) than on the DMP Hub. Can't find a screenshot at this moment.
• One user reports that changing information on contributors can break the citation. Direct quote: "In trying out some changes to the back end, I edited some of the information for the other contributors which led to one now having a blank ORCID link (I never entered any ORCID information for them on the back end), the link to “Investigation” no longer exists for them (it never existed for Crimaldi), and their name no longer appears in the citation. Additionally, it appears that you can add information into affiliation that breaks how it displays if you don’t specifically choose the option from the pop-up menu."
• Sometimes the format splits Data Management || Plan onto a new line
  

This ticket is not to make any changes, just investigate these issues.

Resolving this ticket means:

• We understand how citations are generated
• We can give users instructions on how to enter people in roles/slots to become authors
• We know what other tickets we need to write to resolve any identified bugs (as opposed to just imperfect options that we'll fix in the rebuild)

[bofstein]

Under our new categories I would call this a P2 because it's a minor functionality but affects all (published) users and the wider scientific community. We should at least be able to tell people how the citation is being generated even if each aspect of the fix is not a P2.

[briri]

Both the citation on the plan page within the DMP Tool and the DMP ID landing page, use the JSON metadata stored in the DMP Hub (DynamoDB table).

The citation on the plan page uses this logic (where hash is the DMP record stored I the DMPHub):

    hash = hash.fetch('dmp', {})
    authors = hash.fetch('contributor', []).map { |contrib| contrib['name'] }.join(', ')
    authors = hash.fetch('contact', {})['name'] unless authors.present?
    
    pub_year = Time.parse(hash.fetch('modified', hash['created']))&.strftime('%Y')
    app_name = ApplicationService.application_name

    "#{authors}. (#{pub_year}). \"#{hash['title']}\" [Data Management Plan]. #{app_name}. #{dmp_id}"

So, the plan page will list all contributors OR the owner/creator of the DMP if no contributors exist.

The DMP ID landing page logic looks like this (where investigators is an array of all PI name):

        {investigators.length > 0 &&
          <ul className="landing-list citation">
            <li><strong>When citing this DMP use:</strong></li>
            <li className="margin10 period-separated">
              <span>{investigators}</span>
              <span>({year})</span>
              <span>"{title}"</span>
              <span>[Data Management Plan]</span>
              <span>{dmptoolName}</span>
              <Link href={dmp_id}/>
            </li>
          </ul>
        }

So the landing page displays all of the PIs OR nothing if there are no PIs.

When a change is made in the DMP Tool UI, a background job gets kicked off to update the DMP ID record in the DMP Hub. There is a slight delay before it is synced up. I'm not sure about why it would be losing data as the user described in their comment. It's possible that there is a bug in synchronization process, but this is the first time I'm hearing of it and I'm not able to replicate it. I'm also not sure what the person who reported these issues meant by the "back end" in their comment: "In trying out some changes to the back end".

At the very least I guess we should make them the same.

[bofstein]

My suggestion for current system, IF it's an easy change:

• Citation is all Project Contributors, not just PIs.
• If there are no Project Contributors, then just the owner
• Finalize page and Landing page use same logic
• Order the PIs first, then Data Manager, then Project Administration, then Other
• Edit text on Collaborators page to say for Project Contributors "These are the pe

In future:

• Allow people to check boxes next to both Project Contributors and DMP Collaborators to decide who is included on the citation
• Also allow them to pick the order of it
• Consider given guidance on how to determine the order.