Skip to content

Add GEPA eval upload command#581

Open
willccbb wants to merge 4 commits intomainfrom
codex/add-gepa-upload-support
Open

Add GEPA eval upload command#581
willccbb wants to merge 4 commits intomainfrom
codex/add-gepa-upload-support

Conversation

@willccbb
Copy link
Copy Markdown
Member

@willccbb willccbb commented Apr 29, 2026

Summary

  • Add prime gepa push <run_dir> for uploading Verifiers GEPA outputs as platform evaluations
  • Validate metadata.json and results.jsonl, load optional GEPA artifacts, and preserve GEPA metadata including eval_kind
  • Push samples directly from results.jsonl without injecting task or task_type
  • Keep the existing eval upload flow unchanged

Testing

  • Added unit tests for GEPA run loading, upload request construction, CLI routing, metadata preservation, missing artifact validation, and sample shape handling
  • Ran ruff, ty, and the Prime test suite successfully

Note

Cursor Bugbot is generating a summary for commit 839be1d. Configure here.

@willccbb willccbb requested a review from d42me April 29, 2026 03:40
@willccbb willccbb marked this pull request as ready for review April 29, 2026 03:41
Comment thread packages/prime/src/prime_cli/commands/gepa.py Outdated
Copy link
Copy Markdown

@cursor cursor Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cursor Bugbot has reviewed your changes and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

Fix All in Cursor

❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.

Reviewed by Cursor Bugbot for commit 53f1b56. Configure here.

Comment thread packages/prime/src/prime_cli/commands/gepa.py
@willccbb willccbb requested a review from JannikSt May 6, 2026 22:58
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