define "soft schema freeze" along with exceptions framework

[turbomam]

Especially seeking input from @mslarae13 @aclum @eecavanna @sierra-moxon @shreddd

Steps

• [ ] identify the right place(s) to discuss this and save the decision

[turbomam]

see also

• https://github.com/microbiomedata/nmdc-schema/pull/2032

[eecavanna]

For the exceptions framework (decision framework), I propose this rule:

1. If the PR consists of changes to the nmdc_schema/migrators directory only, then it can be merged in after the typical developer-level review/approval (i.e. without a squad-level discussion/approval).

I expect that situation to arise multiple times during the soft schema freeze.

[aclum]

exception framework proposal:

1. addition of python scripts to nmdc_schema
2. New permissible values for enumerations (ie FileTypeEnum)
3. changes to description/comment slots
4. schema bug fixes identified during migration and testing on existing features

[mslarae13]

We can put the decision in the decision log? There's a note and decision logged in the metadata meeting note

• Schema changes that have any downstream effects will not be permitted after May 31st(June 14th). Downstream relates to requirements for runtime, workflows, products, or any other components of NMDC
• Changes to the schema that do not have downstream effects, are simply modeling changes that can be later implemented can be done on a branch and merged in AFTER review and evaluation.
  • Schema changes that don’t have downstream effects

Decision: Berkeley schema & nmdc-main schema changes should be reviewed and discussed with the team because nmdc-main schema is merged back into berk.

I'm good with @aclum & @eecavanna 's lists for more specific schema changes that don't have downstream effects.

@turbomam where would you like this more "officially" captured?

We also need to make a PR template. Not many people have provided feedback on what that template should look like https://github.com/microbiomedata/nmdc-schema/issues/1995

[turbomam]

@aclum does you exception 1. mean only python changes in nmdc-schema only, not berkeley-schema-fy24? I would want to all changes there too.

Otherwise I agree with the conclusions here.

[aclum]

1. would include python changes in both nmdc-schema, which would then get a back merge to berkeley-schema-fy24, and changes in berkeley-schema-fy24 if need to make them compatible with that version of the schema.

[eecavanna]

We also need to make a PR template. Not many people have provided feedback on what that template should look like https://github.com/microbiomedata/nmdc-schema/issues/1995

Noting that @mslarae13 recently introduced a PR template into berkeley-schema-fy24 via PR https://github.com/microbiomedata/berkeley-schema-fy24/pull/217 and is in the process of introducing one into nmdc-schema via PR https://github.com/microbiomedata/nmdc-schema/pull/2105.