feat: Beginning to flesh out the interface #5
Conversation
* origin/main: chore: Fixing up interface docs: Move docs docs: executors
lewisjared
requested review from
acordonez and
nocollier
and removed request for
lee1043 and
acordonez
|
@lee1043 @acordonez @nocollier @bouweandela Turns out I can only have a single reviewer per PR until this repo is made public... This is broadly how I was thinking about approaching the registration of metrics/providers, i.e each metrics provider has a separate package (and dependencies) inside the repo that implements:
Each metrics provider can then run and test metrics locally independently of the rest of the framework. The ref will then find the appropriate providers and roll that info up. How it does that somewhat depends on how we run things. It might not be as easy as There will be a follow-up RFC to further this out in greater detail. What exactly is part of the inputs/outputs/configuration objects will also expanded on as we need, but should closely align to @acordonez's EMDS standard. |
* origin/main: chore: Always install/choose the local .esgpull directory fix: Typo in the cache name test: Add integration tests chore: Test that regeneration occurs fix: Typo in file to hash docs: Changelog ci: Install esgpull config docs: Update readme chore: Add example of CMEC output from pcmdi ci: Add caching of test data feat: Fetch some example files from esgf
* main: docs: Changelog chore: Fix mypy test: Fix tests chore: Fix mypy feat: validate database urls chore: Update configuration feat: Add sqlalchemy and alembic docs: Changelog chore: Fix tests test: Add tests for the ref package feat: Add a config subcommands feat: Add a ref package
|
I'm going to merge this now so this doesn't block anything. We can update fields etc easily at this stage |
Description
Flesh out a basic interface for a metric.
The goal is to provide an interface that is:
Checklist
Please confirm that this pull request has done the following:
changelog/