At the symposium I spoke to someone from FAIR2ADAPT about a tool to run CODEMETA on a git repository to measure the FAIRness of software. I remember it looked like a useful tool but I have not found (web searching and using AI) a simple description of how to use it with examples. If anyone can point me to the right place that would save me some time and make me less ignorant!
Dear Andy, you talked to me. The tool we discussed is the research software fairness checker: GitHub - oeg-upm/rsfc: Repository for development on the Research Software FAIRness Checks (RSFC) package · GitHub The sample report it produces can be seen at: rsfc/RSFC_REPORT.md at main · oeg-upm/rsfc · GitHub We also have it as a service at https://rsfc.linkeddata.es/ but we are updating our servers now due to security issues, so it may not be up all the time until all issues are solved.
You can use it as an action in your repo (see the readme in the github link above), or as a service with https://rsfc.linkeddata.es/. To use it, just paste a GitHub/GitLab/CodeBerg repo and hit run
Dear @dgarijo , thanks a million for answering my question! I was not sure how to contact you so I posted my question on the forum and it worked - this is the magic of forums! The answer is now there for others to read too.
I am very interested in this tool and the tests it runs. I tried it on one of the critical software packages we develop at the ESRF - GitHub - ewoks-kit/ewoks: Extensible Workflow System · GitHub which is made up of multiple repositories. It is developed and maintained by very experienced research software engineers who take their craft very seriously. I was surprised therefore to see that it passed only 21/40 of the tests. I now have to understand why it did not pass half of the tests which on the surface it should have e.g. RSFC-03-1, RSFC-03-3, RSFC-08-1 and RSFC-18-1. We also have to understand if the missing information needs to be added to each of the git repositories or if we could add it only at the top level. Do you have any guidelines for this use case? I suppose the first step is to create a codemeta file
Anyway thanks again for responding to my post and helping me to find the tool so we can explore it to see how to adopt it.
Hi Andy, the releases one should pass, since I see 43 tags in GitHub. Sometimes GitHub has hiccups, we are working on improving reporting the current status of a repo vs reporting whether we could not run the assessment.We have been running this on over 200 repos in our organization, but we will re-run it in the one you provided to see if there are any problems.
The tool is designed to work/give you guidance per repository basis, so the changes need to occur per repo for now. Different repos usually have different citation files, but it’s an interesting case. How would you propose to provide guidance for multiple repos simultaneously? We can contemplate an extension.
For guiding/generating a codemeta file, we also have https://w3id.org/autocodemeta (but it may be currently down or slow due to the security issues found world wide last week)
I am traveling, but I will get back to you when we re-assess the repo.
Hi @dgarijo , thanks for the feedback and pointing me to the Auto CodeMeta generator. I will use it to complete the missing terms. Have I understood correctly that the terms are all defined and explained here: The CodeMeta Project.
Concerning your question about how to handle a large project which has multiple related repositories I find it would be practical if a single codemeta.json file could cover multiple repositories. This might mean some terms need to be extended to cover multiple repositories.