Skip to content

nemo_retriever: Add structured harness metrics reports#1699

Draft
jioffe502 wants to merge 1 commit intoNVIDIA:mainfrom
jioffe502:fix/harness_metrics
Draft

nemo_retriever: Add structured harness metrics reports#1699
jioffe502 wants to merge 1 commit intoNVIDIA:mainfrom
jioffe502:fix/harness_metrics

Conversation

@jioffe502
Copy link
Collaborator

Switch the retriever harness from stdout scraping to structured run reports shared across supported run modes. This also fixes LanceDB metadata serialization so detection summaries survive real harness runs.

  • add shared batch, inprocess, and fused run-report plumbing and wire the harness to consume it
  • persist runtime and detection metrics as structured artifacts instead of parsing console output
  • tighten coverage around config and LanceDB metadata while relying on real bo20/jp20 harness runs for integration validation

- route harness runs through shared batch, fused, and inprocess reports
- persist runtime and detection metrics without scraping stdout
- tighten tests around config and LanceDB metadata behavior

Signed-off-by: Jacob Ioffe <jioffe@nvidia.com>
@jioffe502 jioffe502 requested a review from a team as a code owner March 23, 2026 21:52
@jioffe502 jioffe502 marked this pull request as draft March 23, 2026 22:00
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