import { createHash } from 'crypto';

/**
 * URL-friendly slug from any source string.
 * Deterministic so the same title always maps to the same URL segment.
 */
export function slugify(input: string): string {
  const slug = input
    .toLowerCase()
    .normalize('NFKD')
    .replace(/[\u0300-\u036f]/g, '') // strip diacritics
    .replace(/^https?:\/\//, '')
    .replace(/[^\w\s-]/g, '')
    .trim()
    .replace(/\s+/g, '-')
    .replace(/-+/g, '-')
    .replace(/^-|-$/g, '')
    .slice(0, 90)
    .replace(/-$/g, '');

  return slug || 'article';
}

/**
 * Stable dedup key for a story title: case-fold, strip punctuation, collapse
 * whitespace. Two feeds covering the same story usually share enough title
 * text to collide here; we add URL-host as a tiebreaker in the pipeline.
 */
export function dedupeKeyOf(title: string, host: string): string {
  const normalized = title
    .toLowerCase()
    .normalize('NFKD')
    .replace(/[\u0300-\u036f]/g, '')
    .replace(/[^a-z0-9\s]/g, ' ')
    .replace(/\s+/g, ' ')
    .trim();
  return createHash('sha1').update(`${normalized}::${host}`).digest('hex').slice(0, 32);
}

export function sha1(input: string): string {
  return createHash('sha1').update(input).digest('hex');
}