github-featured / comet-ml--opik

发布

0 发布 8,340 Git 标签

RSS 订阅源
  • 2.2.3-6110 d94678e1d3
    比较

    [OPIK-7094] [QA] test: cover experiment comparison view (#7550)

    frostbyte_neo 发布于 2026-07-22 10:27:48 +00:00

    Add E2E coverage for the experiment comparison view
    (/experiments/{datasetId}/compare in multi-experiment mode), the gap
    called out in OPIK-7094. Seeds two experiments over one shared dataset
    with divergent per-item outcomes and asserts the side-by-side grid,
    aggregate scores, sort/search, and the per-item drill-in panel.

    • New compare-seed bridge route keeps task_output off the dataset item so
      both experiments run over the same item ids while scoring differently.
    • New comparison fixture + compare-experiments POM + 4-test spec.

    Co-authored-by: Claude Opus 4.8 noreply@anthropic.com

    下载附件