Closed alexlancaster closed 6 months ago
Hi Alex, thanks for making the PR. Instead of merging this one, I'd like to wait for the next version of cffconvert
to be released. In that version the default upload_type
for Zenodo export is software
if users don't include type: dataset
in their CITATION.cff
. I find that logic easier to maintain than having "computed properties".
Hi Alex, thanks for making the PR. Instead of merging this one, I'd like to wait for the next version of
cffconvert
to be released. In that version the defaultupload_type
for Zenodo export issoftware
if users don't includetype: dataset
in theirCITATION.cff
. I find that logic easier to maintain than having "computed properties".
I see, that makes sense. So would that mean that the resource_type: software
would only be added to the related_identifiers
if the upload_type
is set to software
?
I see, that makes sense. So would that mean that the
resource_type: software
would only be added to therelated_identifiers
if theupload_type
is set tosoftware
?
Also, so I know in the future, is there an easier way of updating the unit tests rather than having to go through 40+ .zenodo.json
and Python files to add the new key?
@jspaaks I renamed the branch of my existing work to: https://github.com/alexlancaster/cffconvert/tree/repository_resource_type_zenodo
Let me know if I can retool that branch to work along the lines you suggested above. I can always open it up as a new PR.
Also let me know about how you handle the unit test addition.
According to the CFF Schema: https://github.com/citation-file-format/citation-file-format/blob/develop/schema-guide.md#repository-artifact the two keys:
repository-code
repository-artifact
seem to be intended for software, so this PR adds the
resource_type
to besoftware
in the Zenodorelated_identifiers
to provide more metadata upstream.