/**
 * Content deduplication — detect near-duplicate pages during crawl.
 *
 * Uses MinHash-inspired approach: instead of full MinHash (needs many hash functions),
 * we use 8 structural fingerprints extracted from the markdown to detect pages that
 * share the same layout/template even if the data differs.
 *
 * Typical duplicates: /products?page=1 vs /products?page=2 (same template, different data).
 * These waste crawl budget and generate redundant test cases.
 */
export interface DedupResult {
    isDuplicate: boolean;
    similarUrl?: string;
    similarityScore: number;
}
export declare class ContentDeduplicator {
    private seen;
    private readonly threshold;
    constructor(threshold?: number);
    check(url: string, markdown: string): DedupResult;
    size(): number;
    reset(): void;
}