Two llm_as_judge evaluators score the review generation directly: a
NUMERIC 0-1 on finding actionability, a BOOLEAN on whether the summary
agrees with the findings. Both run on every observation whose trace
name is pr-review or opencode-review.
The judge is kimi-k2.7-code through the headroom hub. Local Ollama
returns Anthropic-format responses but the thinking blocks lack the
signature field Langfuse Zod schema requires; the evaluator preflight
fails as Invalid JSON response. A small judge-proxy pod on 8802
forwards to the hub and patches every thinking block with a synthetic
signature before returning.
Trace + generation output now includes the findings themselves
(capped at 25) rather than just the count, so a judge has something
to grade. generation input/output mirrors the trace so an
observation-level evaluator can read them.
Idempotent: existing evaluators and rules are skipped on re-run,
not duplicated. The connection is upserted on provider.
The filter bar matches on metadata only — not on input and not on the item id —
so a dataset seeded with repo/pr in `input` alone could not be sliced by repo
at all. Every facet worth filtering on is now a flat primitive in `metadata`:
repo, owner, repo_name, pr, head_sha, finding_count, has_findings,
max_severity, reviews_run and the review timestamp both ways. `owner` is split
out because a filter on the joined repo matches one repo, never a whole org,
and `max_severity` is "none" rather than absent because an absent key matches
no filter.
`eval_experiment.py` links reviews that already ran into a dataset run, one run
per model, so the Experiments tab is populated without re-running the reviewer.
One trace per (run, item), the most recent: a PR re-reviewed on every push has
many traces and a run is one output per input.
It posts to the deprecated /api/public/dataset-run-items — the notice exempts
self-hosted v3 from the cutoff date and the pilot is stdlib-only by design.
Revisit at v4.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the evaluation layer on top of the review traces: five deterministic
scores describing how the reviewer behaved, a bridge that turns human reactions
into ground truth, and a dataset seeded from the reviews already run.
The two are kept apart on purpose. feedback.db has recorded 113 reviews and
zero reactions, resolutions or replies — nobody has ever responded to a bot
comment — so an accuracy metric cannot be built yet. The scorers therefore
measure behaviour, which is computable from data in hand, and feedback_scores
turns verdicts into scores the moment any arrive.
eval_scores.py emits finding_rate, severity_info_ratio, severity_max,
dropped_findings and cost_per_finding into the same ingestion batch as the
trace. Undefined values are omitted rather than reported as zero: an info ratio
over a silent review is undefined, and charting it as 0 would read as perfect
calibration.
dropped_findings needed a parser change. Both parsers silently discard findings
with an unusable path/line, which made a model emitting garbage locations
indistinguishable from one that found nothing. last_parse_dropped() exposes the
delta, read at parse time — after apply_repo_config the drops are the config
working as intended, not the model misbehaving.
feedback_scores.py scores the session ("{repo}#{pr}"), because feedback arrives
days later against a PR and nothing records which re-run produced which
comment. review_acceptance is absent rather than 0 when nothing was engaged.
eval_bootstrap.py registers the score configs, seeds the pragent-reviews
dataset, and can backfill scores onto traces that predate the scorers.
expectedOutput is the reviewer's own prior output, flagged
labelled_by_human: false — a regression baseline, not verified truth.
Also fixes a silent telemetry failure: the ingestion endpoint answers 207 when
only some events succeed, so a batch with every event rejected still looked
like success. Score events were missing the required per-event timestamp and
ingested nothing while reporting 207. _warn_on_rejected_events now logs the
per-event errors under LANGFUSE_DEBUG.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude