refactor(evaluations): Remove run mode state

[kaapi-agent-bot]

Issue

Closes #267

Summary

• Removed the Run Mode radio group and its tooltip from RunEvaluationForm.
• Eliminated runMode/setRunMode state and prop threading through EvaluationsTab and /evaluations page.
• Updated create-run payload to no longer include a run_mode field.
• Retained dataset size guard but adjusted warning/error copy to remove "switch to Batch" language.
• Left existing components related to historical evaluation runs untouched.

Checklist

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

• Ran npm run dev and npm run build in the repository root and test.
• If you've fixed a bug or added code that is tested

Original PR description

Summary

RunEvaluationForm let users pick Batch vs. Fast run mode, but v2 evaluation runs are always fast and capped at 500 items (100 original items × 5 duplication factor), so the choice is no longer meaningful.

• Removed the Run Mode radio group and its tooltip from RunEvaluationForm.
• Removed runMode/setRunMode state and prop threading through EvaluationsTab and the /evaluations page.
• The create-run payload no longer sends a run_mode field.
• Kept the existing dataset size guard (eligible_for_fast === false disables the Run button) but reworded the warning/error copy to drop "switch to Batch" language, since Batch is no longer offered.
• Left RunMode, EvalJob.run_mode, and RunModeBadge untouched — they're still used to display the mode of existing/historical evaluation runs (EvalRunCard, /evaluations/[id]), which is unrelated to this issue.

Assumptions

• The dataset_id's eligible_for_fast flag returned by the backend is still meaningful as a general "too large to run" signal even though the mode picker is gone; I kept using it to gate the Run button and to show a size-limit warning, just without mode-specific wording.
• The create-evaluation request no longer needs a run_mode field at all now that fast is the only mode; I removed it from the POST payload rather than hardcoding "fast". Please confirm the backend either ignores a missing run_mode or defaults to fast — happy to send run_mode: "fast" explicitly if that's safer.

Test plan

• npm run lint passes (only pre-existing warnings elsewhere)
• npm run build passes
• Manually verify the Run New Evaluation form no longer shows a Batch/Fast toggle and still runs evaluations end-to-end against a real backend

Closes #267

🤖 Generated with Claude Code

[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: Advanced

Run ID: 25af4bbe-cf3b-4bcb-b2c1-d6345f2ff527

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.}