codemeta validation issues

[mbjones]

I used codemetar 0.1.8 to generate a codemta.json file from the codyn package repository at https://github.com/NCEAS/codyn. I then pasted this codemeta.json file into the CodeMeta Generator at https://codemeta.github.io/codemeta-generator/ and clicked 'Import', which showed the following validation errors, which I fixed one at a time. These problems could either be the codemetar or codemeta-generator not handling things correctly, I'm not sure which.

• @context must be "https://doi.org/10.5063/schema/codemeta-2.0", not ["https://doi.org/10.5063/schema/codemeta-2.0","http://schema.org"]
  • Fixed by removing the schema.org context, but this seems like an overly rigid validator in the codemeta-generator
• "contributor" must be a (list of) Person/Organization object(s) or an URI, but is missing a type/@type.
• `"copyrightHolder" must be a (list of) Person/Organization object(s) or an URI, but is missing a type/@type.
• "funder" must be a (list of) Person/Organization object(s) or an URI, but is missing a type/@type.
• Invalid URL in field "identifier": "10.5063/F1N877Z6"
  • Fixed by turning this into a DOI URL
• Unknown field "sameAs" in "citation".
  • Fixed by removing the element, but this seems like it should work. Need to check the schema.

Several of these are due to being empty elements, so would be fixed by just not generating empty elements:

  "contributor": {},
  "copyrightHolder": {},
  "funder": {},

Getting agreement for the 0.1.9 release between the R tool and the codemeta-generator would help promote adoption. We might want to file an issue at https://github.com/codemeta/codemeta-generator/issues.

[cboettig]

@mbjones thanks for catching and reporting these, all great points.

The @context and the sameAs issue would both be solved if we added http://schema.org into the codemeta schema.

[maelle]

@cboettig I've now changed the milestone to 0.2.0.

Is there any news on the codemeta schema side of things?