/**
 * Evaluation Framework for Drift Detection
 *
 * Runs semantic comparison algorithms against the golden dataset
 * and produces accuracy metrics including precision, recall, F1,
 * and confidence calibration analysis.
 */
import type { EvaluationResult, EvaluationSummary, EvaluationOptions, SemanticComparator } from './types.js';
import { GOLDEN_DATASET, getDatasetStatistics } from './golden-dataset.js';
/**
 * Default semantic comparator using existing implementation.
 */
export declare class DefaultSemanticComparator implements SemanticComparator {
    compare(text1: string, text2: string, toolName: string, category: 'security' | 'limitation' | 'assertion'): {
        matches: boolean;
        confidence: number;
        factors?: Array<{
            name: string;
            weight: number;
            value: number;
            description: string;
        }>;
    };
}
/**
 * Run full evaluation against golden dataset.
 */
export declare function evaluate(options?: EvaluationOptions, comparator?: SemanticComparator): EvaluationResult;
/**
 * Create a summary for display.
 */
export declare function createSummary(result: EvaluationResult): EvaluationSummary;
/**
 * Format evaluation result for console output.
 */
export declare function formatEvaluationReport(result: EvaluationResult): string;
/**
 * Export results as JSON for external analysis.
 */
export declare function exportResultsAsJson(result: EvaluationResult): string;
export { GOLDEN_DATASET, getDatasetStatistics };
//# sourceMappingURL=evaluator.d.ts.map