search

Docmaps-Project / docmaps

Extensible protocol for document history metadata exchange, to enable trustworthy, rapid, open science, by and for preprint science communities.
MIT License
15 stars 1 forks source link

[PROTOCOL]: Renaming a publisher #69

Open giorgiosironi opened 1 year ago

giorgiosironi commented 1 year ago

Protocol semantics improvement

Description

We encountered the case of a publisher name change.

Use case

We consider the publisher of https://sciety.org/docmaps/v1/index?updatedAfter=2023-05-28&publisheraccount=https://sciety.org/groups/rapid-reviews-covid-19 the Sciety group that produced reviews. They are having a big rename from Rapid Reviews COVID-19 to Rapid Reviews Infectious Diseases and the name change has already propagated to the Docmaps we generate.

The docmaps don't have a more recent updated property but where new reviews have appeared we have cases such as duplicated reviews showing up on medrxiv (6 in total rather than the real number, 4).

We haven't touched yet on updating similar publisher information (like the logo) or on something even more groundbreaking, changing their identifier used in publisher.account.id (not even sure that should be allowed).

Proposed solution

While the details of this integration are for us on Sciety (and biorxiv) to sort out, I'm curious as to what we expect to happen to all the old docmaps. Is it useful for them to show the old name if they were produced under that name, or the new name only? In that case do we consider them updated docmaps?

ships commented 1 year ago

Giorgio, you’re right about how much is left unspecified in the docmaps at this time. I expect we will end up with extensions like a key migratedFrom that maps to owl:sameAs or something like that. Alternatively it would be RDFish to allow multiple publisher.account entries though that would be annoying for parsing JSON.

This conversation is to be continued & is being tracked on our internal project board!

ships commented 1 year ago

@giorgiosironi , what is the pathway for medrxiv to discover the set of reviews to display? I'm trying to understand the data model here... did they re-index the docmap with the new publisher and think it was a new review, and appended to a table ?

giorgiosironi commented 1 year ago

Unclear what happened in this specific case, we contacted bioRxiv but the problem was seen as a one-time fix.

In general, they use the API described at https://sciety.org/docmaps/v1 in particular in the https://sciety.org/docmaps/v1/index?updatedAfter=2022-03-25&publisheraccount=https://sciety.org/groups/biophysics-colab for the indexing.