Skip to content

Add pluggable eval metric registry architecture - #3

Closed
AroniasPrenovost wants to merge 2 commits into
boozallen:mainfrom
AroniasPrenovost:feature/eval-metric-registry
Closed

AroniasPrenovost wants to merge 2 commits into
boozallen:mainfrom
AroniasPrenovost:feature/eval-metric-registry

Conversation

@AroniasPrenovost

Copy link
Copy Markdown

Foundational RI-7 roadmap piece: a registry-based interface so new eval metric types (LLM-as-judge, citation accuracy, schema validation, etc.) can be added by implementing EvalMetric and registering it, without touching existing metrics or the eval pipeline core.

Description

Type of change (check as many as applicable)

  • Bug fix (non-breaking change which fixes an issue)
  • Enhancement (non-breaking change which adds functionality)
  • Breaking change (fix or feature that would cause existing functionality to not work as expected)
  • Documentation
  • Maintenance (version updates, config changes, etc)
  • Third-Party/Dependency integration change (schema changes, new/removed/updated APIs)

Work Remaining

How Has This Been Tested?

Pull Request Checklist

  • I have added this PR to the CHANGELOG.md where most applicable
  • I have applied an applicable label to this PR
  • My code follows the style guidelines of this project
  • I have performed a self-review of my own code
  • I have commented my code, particularly in hard-to-understand areas
  • I have made corresponding changes to the documentation
  • I have attached screenshots (if applicable)
  • I have updated the application's SBOM (if applicable)
  • I have updated the DRAFT_RELEASE_NOTES.md (if applicable)
  • I have added tests that prove my fix is effective or that my feature works
  • New and existing unit tests pass locally with my changes
  • Any dependent changes have been merged and published in downstream modules
  • After creating this PR, I have resolved any issues with the automated PR checks

Prenovost, Aron [USA] added 2 commits September 9, 2026 17:02
Foundational RI-7 roadmap piece: a registry-based interface so new eval
metric types (LLM-as-judge, citation accuracy, schema validation, etc.)
can be added by implementing EvalMetric and registering it, without
touching existing metrics or the eval pipeline core.
v3.0.2 is deprecated and GitHub now auto-fails runs that use it.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant