Evaluation

Datasets of (query, expected answer) pairs, runs against a knowledge base in two modes (search retrieval or QA answer), and persisted run history with comparison. Experimental: toggle evaluationDatasets defaults off; enable in Console → Feature Toggles.

Feature	Status	Description
Evaluation dataset CRUD	✅	Create/edit/delete datasets; each linked to one knowledge base
Evaluation items	✅	Add/edit/delete items: query + expected answer pairs; optional topic column; list API paginated (offset/limit, default limit 10); dataset detail UI: per-page size (10/25/50/100), prev/next, range label
CSV import	✅	Import Data button uploads CSV (columns: topic, query, answer or expected_answer)
Run evaluation	✅	POST /api/evaluation-datasets/{id}/run body { evaluation_type }: search_retrieval (default) — hybrid search + LLM judge on snippets; qa_answer — KB QA agent /ask per item + LLM judge on generated answer vs expected; persists evaluation_runs + evaluation_run_items (JSONB detail); response includes run_id, aggregates
Run history & compare	✅	GET .../runs, GET .../runs/{run_id}, DELETE .../runs/{run_id}, GET .../runs/compare?run_a=&run_b=; dataset detail: type selector, history table, load/delete run, compare two runs (per-item pass/score deltas)
Sidebar	✅	"Evaluation" link when evaluationDatasets toggle enabled
Feature toggle	✅	evaluationDatasets (default: false); Console → Feature Toggles