Closed joverlee521 closed 1 month ago
I'm a bit out of the loop here. What's the difference between mpox's CI (which uses pathogen-repo-build
) and avian-flu's CI (which uses pathogen-repo-ci
)? Can avian-flu use pathogen-repo-build
? Or is mpox's use of pathogen-repo-build
considered temporary and will be replaced during the 3rd task here?
Or is mpox's use of pathogen-repo-build considered temporary and will be replaced during the 3rd task here?
Yup, the use of the pathogen-repo-build
in CI is considered a temporary change (if ~6 months is considered temporary 😅). The plan would be to switch all pathogen GitHub Action CI workflows back to pathogen-repo-ci
.
Related/precursor work: https://github.com/nextstrain/.github/issues/89
Ideas here came out of discussion with @tsibley in our 1:1 regarding https://github.com/nextstrain/.github/pull/62#issuecomment-2105402378. This would supersede #63 + #62.
Proposal
With the standardization of pathogen repos following the pathogen-repo-guide, we can now enforce the standardized use of the
pathogen-repo-ci
workflow.In order to use the centralized
pathogen-repo-ci
, a pathogen repo will need to meet two requirements.phylogenetic
subdirectory)use a standard mechanism for specifying the CI build, e.g. use a standard CI config that can be specified with
This would simplify each pathogen repo's CI workflow to:
Then the docker-base/conda-base CIs can be simplified to just matrix of pathogen repos instead of needing to pass in additional build args.
Concrete steps