So glad to see this work being done! Thanks again @MartinPaulEve.

I am looking forward to testing out the interface. I wasn't able to do so yet because I ran into a migration error:

CommandError: Conflicting migrations detected; multiple leaf nodes in the migration graph: (0083_creditrecord_preprint_author, 0085_alter_article_jats_article_type_override_and_more in submission).

What we like to do nowadays with git conflicts or split migration graphs is rebase dev branches on master and change the dependency of the dev migration. This allows us to avoid creating duplicate merge migrations in multiple people's dev instances. Could you do that please before I go further with my review?

Thanks again for this Martin.

I wrote some changes inline that I'd like to see before merging. Some of them are important to take care of regardless of how the UI works, while some of them could be informed by the conversation we have around #4519 . More details inline.

Two global comments:

• Shouldn't there be an interface for adding and removing CREDIT roles in the repository submission system, if there is a field for repository author on the role object? Or did I miss that integration?
• Shouldn't the canonical URI to each CRediT role be captured to make it more machine-readable in your plugin and other places like our own JATS? Taking a quick look at the implementation guidelines, it seems like the URLs would be useful for name authority. I think this is especially the case since the strings used contain questionable characters for downstream consumption, like the ampersand in "Writing - Review & Editing".

Hi @MartinPaulEve, thanks for this progress.

I still think we need to think more about how the snapshot_self function works with CRediT roles. What about this scenario: An author submits an article on which they are drafting author, and then, while it's in production, they submit another one to the same journal on which they are the editing author. How does Janeway not erroneously pick up the editing author role from article 2 when it's updating the frozen authors for article 1, say in typesetting or prepub? This is a harder problem to solve, because I think it might involve creating frozen authors when author metadata is first entered during submission, and not updating them when we snapshot the author at a later date, just allow editing the frozen author. That's perhaps beyond the scope of your work, but I want to raise it here so @mauromsl and @ajrbyers are aware of it, and we'll have to address it before we release 1.8.

the new interface ... should simply hide the ability to add CRediT information if it's disabled

Yeah that's what I was thinking, but how does an editor or press manager disable it? Have I missed a setting you have introduced? Any thoughts on my suggestion above: "A boolean-type core.models.Setting should do the trick, and you can add one in src/utils/install/journal_defaults.json." The setting should be introduced in core Janeway before you try to make your OA switchboard plugin hook into it, because otherwise you won't know how to check for whether CRediT roles are expected or not, for a given journal's metadata. Ideally, this setting, not whether CRediT records are present, should control whether this metadata is displayed on the article page. Editors will expect a single toggle on / off to control it everywhere.

so I've added CRediT to the JATS, including the work on canonical URLs (which Mauro suggested to store in the choices list, rather than in the database)

Super, though I'm wondering why we have spaces and ampersands in the DB keys? Any reason why we shouldn't use the less error-prone short slug form of each role, e.g. formal-analysis? We can also change this after this PR is merged if you don't have time to keep making changes like this.

Again, really appreciate this contribution. As I have upcoming annual leave I am going to request changes and also request Andy's and Mauro's review so you have some other perspectives sooner rather than later.