Crate runifold_testkit 

Deterministic helpers for testing Runifold runtime semantics.

Structs

EvaluationCase

One immutable evaluation input and its optional reference answer.

EvaluationCaseId

Stable case identity within a dataset.

EvaluationCaseResult

Output-free per-case evaluation result.

EvaluationDataset

Versioned, duplicate-free evaluation dataset.

EvaluationFailure

Safe per-case execution or scorer failure.

EvaluationMetrics

Non-sensitive resource evidence for one successful target execution.

EvaluationOutput

Target output plus optional Run/Trace correlation.

EvaluationReport

Deterministic candidate evaluation report.

EvaluationRunner

Concurrent deterministic evaluation orchestrator.

EvaluationScore

One persisted per-case score.

EvaluationScoreSummary

Aggregate score statistics.

FileEvaluationRepository

Filesystem-backed immutable JSON evaluation repository.

FnScorer

Closure-backed asynchronous scorer.

JsonExactMatchScorer

Deterministic JSON equality scorer.

JsonRuleScorer

Weighted deterministic scorer for structured outputs.

JudgeRubric

Versioned rubric for a structured model judge.

MetricRegression

One baseline-to-candidate score comparison.

ModelJudgeScorer

Canonical-model-backed structured LLM judge.

ModelScenario

Accumulates a scripted provider event sequence without network access.

RegressionComparison

Complete relative regression decision.

RegressionPolicy

Allowed relative quality drops.

RunScenario

An isolated root run with deterministic event collection.

ScoreValue

Validated score value and optional evaluator rationale.

ScriptedModel

A deterministic, queue-backed model adapter.

TokenOverlapScorer

Deterministic lexical overlap scorer for string reference answers.

WeightedJsonRule

One weighted deterministic JSON-output rule.

Enums

EvaluationError

Evaluation configuration or execution failure.

EvaluationFailureStage

Evaluation failure stage.

EvaluationStoreError

Evaluation artifact repository failure.

JsonRule

Deterministic rule over a target JSON output.

Traits

EvaluationRepository

Versioned dataset and candidate-report repository.

EvaluationScorer

Asynchronous scorer boundary.

EvaluationTarget

Asynchronous system-under-evaluation boundary.

Type Aliases

EvaluationFuture

Owned asynchronous evaluation operation.