enhancement(evaluation): Reference summary findings by question number

[AkhileshNegi]

Issue

Closes #1174

Summary

The evaluation run summary keyed each per-question judge trace on the Langfuse dataset item id (item_{row}_{dup}) and told the summary model to cite that id. Reviewers got a "Top 3 to check" list of item_3_1 / item_12_4, which cannot be mapped back to a question without opening Langfuse — and with duplication factor > 1 the same question appears under several ids, so one bad question could eat all three slots.

• Summary brief now keys on question_id (app/crud/evaluations/summary.py): _format_traces_for_prompt hands over the 1-based dataset row number instead of trace_id, and the system prompt asks the model to reference each question as "Question N". question_id was already populated end to end (langfuse.py writes it into trace metadata, merge.py preserves and sorts on it), so this is a consumer-side change only.
• Stronger judge model: EVAL_JUDGE_MODEL moves from gpt-5-mini to gpt-5.6-luna (app/core/config.py).
• Tests updated for the new payload key (app/tests/crud/evaluations/test_run_ai_summary.py).

Checklist

Before submitting a pull request, please ensure that you mark these task.

• Ran fastapi run --reload app/main.py or docker compose up in the repository root and test.
• If you've fixed a bug or added code that is tested and has test cases.

Notes

The ground-truth judge prompt rewrite that was previously on this branch has been reverted — this PR is now scoped to the summary id change and the judge model bump.

[coderabbitai]

Important

Review skipped

Auto reviews are limited based on label configuration.

🏷️ Required labels (at least one) (1)

• ready-for-review

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: c71312d0-10f5-4078-a3a5-095f9305e1f8

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

OpenAPI changes   ⚪ No API surface changes

Note

This PR does not modify the API contract.

_{main ↔ 576fd3a7 · generated by oasdiff}