[NA] [FE] fix: deduplicate eval suite input columns in experiment items table - #6181
Conversation
…ms table Filter out output columns whose names already appear in dataset columns to avoid showing duplicate data (e.g. "input") in both the "Evaluation suite" and "Evaluation task (last trial)" table sections. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Keep dynamicOutputColumns intact in the shared hook so filters and column visibility state are unaffected. Apply the dedup filter only when building outputColumnsData for table rendering. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
![baz-reviewer[bot]](https://avatars.githubusercontent.com/in/933528?s=60&v=4){kind=small}
… filters, and column picker Compute visibleOutputColumns once and use it for outputColumnsData, filterColumns, and columnSections so deduped columns don't appear as phantom entries in the filter panel or column picker. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
… only Only hide the "input" output column in eval suite mode, where the evaluation task echoes back the dataset input. Regular experiments are completely unaffected. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
🔄 Test environment deployment process has started Phase 1: Deploying base version You can monitor the progress here. |
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
✅ Test environment is now available! To configure additional Environment variables for your environment, run [Deploy Opik AdHoc Environment workflow] (https://github.com/comet-ml/comet-deployment/actions/workflows/deploy_opik_adhoc_env.yaml) Access Information
The deployment has completed successfully and the version has been verified. |
|
🌙 Nightly cleanup: The test environment for this PR ( |
|
🌙 Nightly cleanup: The test environment for this PR ( |
1 similar comment
|
🌙 Nightly cleanup: The test environment for this PR ( |
…ms table (#6181) * [NA] [FE] fix: deduplicate eval suite input columns in experiment items table Filter out output columns whose names already appear in dataset columns to avoid showing duplicate data (e.g. "input") in both the "Evaluation suite" and "Evaluation task (last trial)" table sections. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(experiments): move column dedup to table display layer only Keep dynamicOutputColumns intact in the shared hook so filters and column visibility state are unaffected. Apply the dedup filter only when building outputColumnsData for table rendering. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(experiments): use consistent deduped output columns across table, filters, and column picker Compute visibleOutputColumns once and use it for outputColumnsData, filterColumns, and columnSections so deduped columns don't appear as phantom entries in the filter panel or column picker. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(experiments): scope eval suite output dedup to echoed "input" key only Only hide the "input" output column in eval suite mode, where the evaluation task echoes back the dataset input. Regular experiments are completely unaffected. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(experiments): fix prettier formatting and implicit any types Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
3 participants
Details
When viewing evaluation suite experiments, the "input" column appeared twice — once under "Evaluation suite" (from dataset item data) and again under "Evaluation task (last trial)" (from experiment item output). This happened because the SDK always produces
{"input": ..., "output": ...}as the task result, so the "input" key in the output always overlaps with the dataset's "input" key.This fix filters
dynamicOutputColumnsto exclude the second "input" column when displaying the experiment item for evaluation suites.Change checklist
Issues
AI-WATERMARK
AI-WATERMARK: yes
Testing
npx tsc --noEmit) — passesDocumentation
N/A