Adding contributor roles to the metadata

[dhimmel]

@nicolevasilevsky wrote:

Add something here about using CRO? Manubot should be be able to track our contributions in the yaml file too?

Yes, we can put extra fields for authors in metadata.yaml. For example we could do:

  - github: nicolevasilevsky
    name: Nicole Vasilevsky
    initials: NV
    orcid: 0000-0001-5208-3432
    twitter: n_vasilevsky
    email: vasilevs@ohsu.edu
    affiliations:
      - Oregon Clinical and Translational Reseearch Institute, Oregon Health & Science University
      - Department of Medical Informatics and Clinical Epidemiology, Oregon Health & Science University
    roles:
      - CREDIT_00000013 writing original draft role

Currently we don't do anything with that information, but we could. @nicolevasilevsky should we show the roles in the manuscript?

[nicolevasilevsky]

I like the idea of showing the roles in the manuscript, thanks for doing that!

cc @kristiholmes

[nicolevasilevsky]

I will add the roles to the yaml file

[dhimmel]

Okay I'll touch base with our frontend dev @vincerubinetti on what might be some nice ways to display these roles in the manuscript frontmatter.

[nicolevasilevsky]

awesome, thanks!

[kristiholmes]

Agree - this is awesome! Thanks!

[jcolomb]

Please have a look at : https://jats4r.org/credit-taxonomy:
it would be best to define exactly what is under the role, url seem prefered to the PID ?

We should all discuss how to include CRO information.

[mellybelly]

@jcolomb the credit-taxonomy was incorporated into CRO; we were part of the original workshops that defined the taxonomy. CRO is a computational representation and a much more extensive set of concepts, however the original credit taxonomy can still be used as a subset.

[jcolomb]

the question is how to implement the CRO vocabulary into the paper metadata, see the link: there is a recommendation when using credit, but not for CRO.

[jcolomb]

An additional thought: any way to add contributors who are not authors something like:

- github: nicolevasilevsky
    name: Nicole Vasilevsky
    initials: NV
    orcid: 0000-0001-5208-3432
    twitter: n_vasilevsky
    email: vasilevs@ohsu.edu
    affiliations:
      - Oregon Clinical and Translational Reseearch Institute, Oregon Health & Science University
      - Department of Medical Informatics and Clinical Epidemiology, Oregon Health & Science University
    roles:
      - CREDIT_00000013 writing original draft role
    is_author: yes
    is_corresponding_author: no

(PS: I moved question about contributor role in #8)

[dhimmel]

any way to add contributors who are not authors something like

Do you want the translators to show up on the manuscript, like we did for translators at https://greenelab.github.io/scihub-manuscript-es/#traductores?

For now we could just create another field in the YAML for contributors below authors if we have this situation. Then I can deal with making it show up in the manuscript later.

[dhimmel]

Please have a look at : https://jats4r.org/credit-taxonomy:
it would be best to define exactly what is under the role, url seem prefered to the PID ?

I like the URLs for their human-readability and the ability to be resolved in a web browser, but the URLs like https://dictionary.casrai.org/Contributor_Roles/Visualization return "Oops! That page can’t be found."

I was thinking we could just put - CREDIT_00000013 writing original draft role for now and then we can convert this to a different format (or ignore everything but the ID) when we decide how we want to process this and include it in manuscript.

[mellybelly]

This is why we created this resolvable version in OWL: https://github.com/data2health/credit-ontology
see also http://obofoundry.org/ontology/cro.html
which imports the original CREDIT terms.

(note that some of the original terms are a bit conflated, so this is a bit of a problem)

[mellybelly]

@dhimmel @cgreene it would be terrific if manubot could try to implement the CAM model and CRO :-)

[cgreene]

I think I tweeted in the past that I would love to see author contributions on mouseover (or similar) in author lists.

[jcolomb]

concerning contributor role description, I would take these actions :

• get the cd2h group to define the best way to mention contribution in publication metadata (get a PID of CRO content?), make sure it is compatible with what people did so far.
• make the jats4r change their recommendation accordingly
• implement a yaml version of this
• define manubot output from the yaml

Make sense ?
(PS: I moved question about contributor role in #8)

[jcolomb]

this is what I can get out of my program fetching orcid info:
(I started with xml export, so the names of the categories are different, can we work with this?)

name:
  given-names: Julien
  surname: Colomb
contrib-id: 0000-0001-9799-2656
github-handle: https://github.com/jcolomb
twitter-handle: []
author-notes:
  email: julien.colomb@fu-berlin.de
affiliation:
- Humboldt-Universität zu Berlin, SFB1315, Institut für Biologie, Charitéplatz 1,
  10117 Berlin
- Friedrich-Schiller-Universität Jena, Informatik  (RDM helpdesk)
role:
- Conceptualization
- Software
funders:
- 'funded by the German Research Foundation, grant_number 327654276: SFB 1315:             Mechanisms
  and disturbances in memory consolidation: From synapses to systems'
- 'funded by the Swiss National Science Foundation, grant_number grant.116951: Memory
  phases of reward learning in Drosophila melanogaster'
.attrs:
  contrib-type: author
  corresponding-author: no

[dhimmel]

concerning contributor role description, I would take these actions :

I agree with those steps, I think. For now, I think we should continue adding authors roles to metadata.yaml in whatever format and then we can standardize at a later point, as to not hold things up.

get the cd2h group to define the best way to mention contribution in publication metadata (get a PID of CRO content?)

BTW, I noticed CRO wasn't on https://registry.identifiers.org/registry. This service is a great way to get a PID that can be interpreted by as many applications as possible.

@jcolomb I am treating this issue as "let's add contributor roles to this specific manuscript", and not necessarily, "let's solve the problem forever". I would like a good implementation, but I view this as more proof of concept. Happy to work on more general / permanent solutions & standards for Manubot. I wasn't clear whether you were making suggestions for this manuscript specifically or for what we should do with more time going forward, especially if it depends on third-party changes.

I started with xml export

An XML export of what? This manuscript using Pandoc to create JATS?

this is what I can get out of my program fetching orcid info

This looks cool. I want to mention manubot/manubot#187 where we're discussing what fields metadata.yaml should have generally. What is your program, and is this relevant to this specific manuscript, or something you'd like to integrate in the future?

[mellybelly]

Hi @dhimmel
that is strange RE identifiers.org; all OBO prefixes should by default be there. I will make a ticket on the OBO repo.

We also like to ensure multiple resolvers in all cases, and use https://n2t.net/
They have a nice prefix management system.

In the https://github.com/prefixcommons project we aimed to coordinate prefixes across the two, but it doesn't have funding at present and perhaps not everyone is synched at present.

@cmungall @jmcmurry can you advise?

[jcolomb]

Since a part of the paper is about how to implement contributor roles annotation in the literature, the two (this paper and long term solutions) are linked to some extend.
Maybe we should try to:

• do the best we can on a short term basis for this paper
• document what is needed to do in the future to integrate in the paper discussion
• keep in touch for the implementation of these long term solution (in manubot and beyond)

Should we move that discussion to slack?

[dhimmel]

@jcolomb makes sense. I'm happy to work together to do as much as possible for this paper and beyond.

Should we move that discussion to slack?

I prefer github issues for being public and integrated with the source code. Feel free to make new issues anywhere and tag me. Just saw your open-science-promoters/contibutor_manager repo... looks cool.

This is what I'm thinking for the general implementation:

1. contributor roles are defined in the the metadata, using whatever standard we decide is best
2. a script / program retrieves the name and description (examples). Ideally, there'd also be badges / logos for each role, but that probably doesn't exist.
3. We add these roles to the outputs as structured metadata (however is appropriate)
4. We add these roles to the HTML output in such a way that you can see all the roles (including extended definitions of those roles) for each author. We probably want to use some sort of tool tip or expandable element as to not overwhelm viewers who are not interested.

Does this make sense @jcolomb ?

[jcolomb]

Discussion moved to manubot repository: manubot/manubot#202