export type BenchmarkMode = "baseline" | "devflow";
export type BenchmarkScenarioCategory = "understand" | "modify" | "review";
export interface BenchmarkTask {
    id: string;
    type: "bugfix" | "feature" | "review";
    desc: string;
    expectedFiles?: number;
    expectedEndpoints?: number;
    expectedComponents?: number;
    expectedIssues?: number;
}
export interface TaskResult {
    taskId: string;
    baseline: {
        tokens: number;
        time: number;
        correct: boolean;
    };
    experiment: {
        tokens: number;
        time: number;
        correct: boolean;
    };
    delta: {
        tokenReduction: number;
        timeSaved: number;
        accuracy: number;
    };
}
export interface BenchmarkReport {
    runId: string;
    timestamp: number;
    summary: {
        avgTokenReduction: string;
        avgTimeSaved: string;
        avgAccuracyGain: string;
        avgConfidenceGain: string;
    };
    tasks: TaskResult[];
}
export interface ReportMeta {
    runId: string;
    timestamp: number;
    taskCount: number;
}
export interface BenchmarkScenario {
    id: string;
    category: "understand" | "modify" | "review";
    desc: string;
    expectedFiles?: number;
    expectedEndpoints?: number;
    expectedComponents?: number;
    expectedIssues?: number;
}
export interface BenchmarkRunResult {
    scenarioId: string;
    mode: BenchmarkMode;
    prepTimeMs: number;
    tokenCount: number;
    firstPassSuccess: boolean;
    reworkCount: number;
    durationMs: number;
    notes?: string;
    timestamp: number;
}
export interface ComparisonResult {
    timeSavedRate: number;
    tokenReductionRate: number;
    firstPassLift: number;
    reworkReductionRate: number;
}
export interface BenchmarkComparisonResult {
    scenarioId: string;
    baseline: BenchmarkRunResult;
    devflow: BenchmarkRunResult;
    comparison: ComparisonResult;
}
export type BenchmarkMetricName = "tokensUsed" | "filesTouched" | "correctnessScore" | "durationMs";
export interface BenchmarkSuiteTask {
    id: string;
    prompt: string;
    keywords: string[];
    expectedFiles: string[];
    expectedSymbols?: string[];
    scoringDimensions: BenchmarkMetricName[];
    sourceAccuracyQuery?: string;
}
export interface BenchmarkSuite {
    id: string;
    name: string;
    version: string;
    description: string;
    tasks: BenchmarkSuiteTask[];
}
export interface BenchmarkLegMetrics {
    mode: BenchmarkMode;
    tokensUsed: number;
    filesTouched: string[];
    correctnessScore: number;
    durationMs: number;
    matchedFiles: string[];
    matchedSymbols: string[];
    rounds: number;
}
export interface BenchmarkTaskComparison {
    taskId: string;
    prompt: string;
    baseline: BenchmarkLegMetrics;
    devflow: BenchmarkLegMetrics;
    delta: {
        tokenReductionRate: number;
        fileReductionRate: number;
        correctnessLift: number;
        timeSavedRate: number;
    };
    evaluation?: BenchmarkTaskEvaluation;
}
export interface BenchmarkGraderResult {
    graderId: string;
    kind: 'deterministic' | 'model' | 'human';
    status: 'passed' | 'failed' | 'unavailable';
    score?: number;
    threshold?: number;
    evidence: Array<{
        type: string;
        value: string;
    }>;
    reason: string;
    durationMs: number;
    receipt: string;
}
export interface BenchmarkTaskEvaluation {
    status: 'passed' | 'failed' | 'unavailable';
    results: BenchmarkGraderResult[];
    blockingFailures: string[];
    unavailableRequired: string[];
}
export interface BenchmarkAggregateMetrics {
    avgTokensUsed: number;
    avgFilesTouched: number;
    avgCorrectnessScore: number;
    avgDurationMs: number;
}
export interface BenchmarkRegression {
    metric: BenchmarkMetricName;
    changeRate: number;
    threshold: number;
    message: string;
}
export interface BenchmarkSuiteReport {
    runId: string;
    suiteId: string;
    suiteName: string;
    suiteVersion: string;
    projectRoot: string;
    commit: string | null;
    createdAt: number;
    rounds: number;
    status: "completed";
    tasks: BenchmarkTaskComparison[];
    summary: {
        baseline: BenchmarkAggregateMetrics;
        devflow: BenchmarkAggregateMetrics;
        delta: {
            tokenReductionRate: number;
            fileReductionRate: number;
            correctnessLift: number;
            timeSavedRate: number;
        };
    };
    regressions: BenchmarkRegression[];
    evaluation?: {
        status: 'passed' | 'failed' | 'unavailable';
        passedTasks: number;
        failedTasks: number;
        unavailableTasks: number;
    };
}
export interface BenchmarkReportMeta {
    runId: string;
    suiteId: string;
    suiteVersion: string;
    projectRoot: string;
    commit: string | null;
    taskCount: number;
    createdAt: number;
    status: "completed";
}
//# sourceMappingURL=benchmark.d.ts.map