microbiomedata/nmdc-schema

`berkeley`: Incorporate `refscan` graphs into schema documentation

Closed this issue · 8 comments

Idea: Use GitHub Actions to install refscan from PyPI, run it on the latest schema, and store the resulting HTML file in the right place in the schema docs so that it appears on the deployed schema docs website.

Emphasis: This does not involve adding any more "derived files" to the repo. The docs are built by GHA.

Limitation: refscan only works with the Berkeley schema (due to refscan's dependence upon the type field, which is only "reliable" in the Berkeley schema).

During today's NMDC sync, people seemed to me to want this to be a public-facing diagram and some of them seemed to me to be in favor of it living in the schema repo (in this case, it would start out in the berkeley-schema-fy24 repo and only appear in the nmdc-schema repo once the fork gets merged into the base). Needs buy-in from @turbomam, specifically, as lead maintainer of the repo.

Currently, graphs are hosted at this improvised URL: https://portal.nersc.gov/project/m3408/refscan/

I support adding the refgraphs to the nmdc-schema documentation. I will probably have some suggestions for brief documentation that could go along with it.

I support adding the refgraphs to the nmdc-schema documentation.

Thanks, @turbomam. I'll try to make this happen next sprint.

I will probably have some suggestions for brief documentation that could go along with it.

Sounds good to me.

I've prototyped this in berkeley-schema-fy24 PR microbiomedata#265.

@turbomam @sujaypatil96 can you provide feedback on PR microbiomedata#265 please?

@eecavanna I'm guessing this is on hold till after the Berkeley rollout correct? Can we move this to next sprint?

Yes, next sprint, please. We did decide to hold off on merging this until after the Berkeley Schema Roll Out (even if the PR has been reviewed/approved before then).

This has been implemented in #2198 and I am ready for it to be merged in.

Part of this has been re-implemented (more granularly) in PR #2240.

Closing this ticket in favor of the following two, newer, more narrowly-scoped tickets: