feat: add Crossref DOI provider

[mfenner]

• register DOIs with Crossref, functionality analogous to DataCitePIDProvider
• CrossrefXMLSerializer with metadata export

❤️ Thank you for your contribution!

Description

This pull-request adds Crossref DOI registration support to InvenioRDM. The configuration is very similar to DataCite DOI registration, and the new PID provider should not interfere with DataCite DOI registration.

The Crossref PID provider makes heavy use of the commonmeta-py library, similar to how the datacite library is used in the DataCite PID provider. commonmeta-py already was a dependency, but a newer version (>= 0.183) is required for Crossref DOI registration functionality.

Checklist

Ticks in all boxes and 🟢 on all GitHub actions status checks are required to merge:

• I'm aware of the code of conduct.
• I've created logical separate commits and followed the commit message format.
• I've added relevant test cases.
• I've added relevant documentation.
• I've marked translation strings.
• I've identified the copyright holder(s) and updated copyright headers for touched files (>15 lines contributions).
• I've NOT included third-party code (copy/pasted source code or new dependencies).
  • If you have added third-party code (copy/pasted or new dependencies), please reach out to an architect on Discord.

Frontend

• I've followed the CSS/JS and React guidelines.
• I've followed the web accessibility guidelines.
• I've followed the user interface guidelines.

Reminder

By using GitHub, you have already agreed to the GitHub’s Terms of Service including that:

1. You license your contribution under the same terms as the current repository’s license.
2. You agree that you have the right to license your contribution under the current repository’s license.

[mfenner]

Opened this PR instead of #2177 as the PR was open for too long.

[tmorrell]

Looks like a nice well-tested addition. I have one question about the vocabulary changes that should be addressed, but otherwise looks good to go

[mfenner]

Sorry for the extra commits, I accidentally named one test file with a capital T, and MacOS doesn't care. And there are issues with tests that I need to again look into.

[mfenner]

I have fixed most tests and added one new feature. I will do more cleanup via rebasing.

Aware that the Invenio partner meeting is less than three weeks away and that this is the best chance to get this PR discussed and approved, I added the following functionality:

• Allow Crossref and DataCite DOI registration in the same instance
• Allow multiple DOI prefixed per provider in addition to the default prefix both for Crossref and DataCite
• Control the DOI Provider and DOI prefix in the REST API call when creating the record

All of the above is optional, the DataCite defaults have not changed, and Crossref is (almost) a drop in replacement, i.e. all DOIs are registered via Crossref and with one prefix. I say almost, as there are some subtle differences.

I left out the decision flow and UI part of how a DOI provider and prefix are picked if there are multiple. This is tricky and needs more discussion. There is strong interest in one prefix per community, but records can be registered before a community is added and moved to a different community. The other major decision point is content type, as Crossref only supports text content types. But the UI of changing the content type of a record then becomes more complicated. A good discussion topic at the workshop.

This PR can be merged before or at the workshop. It provides a good foundation for more flexibility in DOI registration in v14. Instances can build on that addressing their use cases, or we might have a good discussion at the workshop about a first UI implementation. I will deploy the branch (or the merged master) in my production instance this week, so that I can test DOI registration for a few hundred DOIs (my usual quota) until the workshop.

[mfenner]

I fixed all tests - the Python 3.9 tests fail because the commonmeta-py dependency requires 3.10 in recent versions. Also seing no issues in my production instance so far, with a few hundred records registered with Crossref with this latest version.

[utnapischtim]

@mfenner could it be that you missed to add the CROSSREF_ prefix to be loaded in ext.py like datacite here

[mfenner]

@utnapischtim what would this enable? I am registering DOIs with Crossref in my instance without issues.

[utnapischtim]

@mfenner
yes, and i think you will configure the crossref configuration variables in your invenio.cfg file (or somewhere else) but if you set up a new instance and you don't configure CROSSREF_* in your invenio.cfg file, no crossref configuration variables would be loaded. so this would fail with

invenio_rdm_records/services/pids/providers/external.py", line 32, in prefixes
    val = current_app.config[name]
          ~~~~~~~~~~~~~~~~~~^^^^^^
KeyError: 'CROSSREF_PREFIX'

Loading the values as i suggested would fix the problem.

[mfenner]

@utnapischtim added CROSSREF_ prefix as you suggested.

[mfenner]

@utnapischtim I merged the latest code from master, included a new commonmeta-py version (0.215) which uses a simpler approach to crossref_xml generation and supports Python 3.9, and fixed the failing crossref serializer test. Tried to rebase, but still not comfortable doing this, especially in such an important package. All tests. pass now and I will continue to use the branch in my production instance.