import 'server-only';
import { api, safe } from '@/lib/api';
import { routes, SITE_URL, TYPE_PATH } from '@/lib/site';

export const SITEMAP_HEADERS = { 'content-type': 'application/xml; charset=utf-8', 'cache-control': 'public, s-maxage=3600, stale-while-revalidate=86400' };
export const SHARD = 5000;
/** Static public routes (1.1 surface). Query-driven pages are listed once at their canonical default URL. */
const STATIC = [
  '/',
  '/models',
  '/companies',
  '/papers',
  '/providers',
  '/benchmarks',
  '/benchmarks?view=matrix',
  '/hardware',
  '/hardware/fit',
  '/frameworks',
  '/datasets',
  '/tools',
  '/agents',
  '/families',
  '/licenses',
  '/open',
  '/frontier',
  '/prices',
  '/pulse',
  '/calculator',
  '/run-locally',
  '/find-a-model',
  '/compare',
  '/graph',
  '/changes',
  '/timeline',
  '/time-machine',
  '/diff',
  '/explore',
  '/methodology',
  '/sources',
  '/about',
  '/developers',
  '/bot',
];
/** Types with their own sitemap series (in this order); everything else goes into the "other" series. */
const SERIES = ['model', 'company', 'paper', 'provider', 'benchmark', 'hardware', 'framework', 'dataset', 'tool', 'repository', 'model_family', 'artifact', 'license'] as const;

function esc(s: string): string {
  return s.replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/"/g, '&quot;');
}

export function toIndex(locs: string[]): string {
  return `<?xml version="1.0" encoding="UTF-8"?>\n<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n${locs.map((l) => `  <sitemap><loc>${esc(SITE_URL + l)}</loc></sitemap>`).join('\n')}\n</sitemapindex>\n`;
}
export function toUrlset(entries: { loc: string; lastmod?: string }[]): string {
  return `<?xml version="1.0" encoding="UTF-8"?>\n<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n${entries.map((e) => `  <url><loc>${esc(SITE_URL + e.loc)}</loc>${e.lastmod ? `<lastmod>${e.lastmod.slice(0, 10)}</lastmod>` : ''}</url>`).join('\n')}\n</urlset>\n`;
}

/** Shard ids: `static`, `changes` (daily digests of the observation history), then `<type>-<n>` per type with entities. */
export async function shardIds(): Promise<string[]> {
  const ids = ['static', 'changes'];
  const stats = await safe(api.stats());
  for (const t of SERIES) {
    const n = Number(stats?.entities?.[t] ?? 0);
    const shards = n > 0 ? Math.ceil(n / SHARD) : 0;
    for (let i = 0; i < shards; i++) ids.push(`${t}-${i}`);
  }
  return ids;
}

/** One /changes/<date> URL per UTC day from the first entity to today (bounded to 3 years). */
async function changesEntries(): Promise<{ loc: string; lastmod?: string }[]> {
  const stats = await safe(api.stats());
  const first = stats?.first_entity_at ? new Date(stats.first_entity_at) : null;
  if (!first || Number.isNaN(first.getTime())) return [];
  const today = new Date();
  const start = Math.max(first.getTime(), today.getTime() - 3 * 365 * 86400000);
  const out: { loc: string; lastmod?: string }[] = [];
  for (let t = start; t <= today.getTime(); t += 86400000) {
    const day = new Date(t).toISOString().slice(0, 10);
    out.push({ loc: routes.changesDay(day) });
  }
  return out;
}

export async function shardEntries(id: string): Promise<{ loc: string; lastmod?: string }[] | null> {
  if (id === 'static') return STATIC.map((loc) => ({ loc }));
  if (id === 'changes') return changesEntries();
  const m = /^([a-z_]+)-(\d+)$/.exec(id);
  if (!m) return null;
  const type = m[1] as string;
  const n = Number(m[2]);
  if (!(SERIES as readonly string[]).includes(type) || !TYPE_PATH[type]) return null;
  const res = await safe(api.sitemap(type, SHARD, n * SHARD));
  if (!res) return [];
  return res.items.map((it) => ({ loc: routes.entity({ entity_type: it.entity_type, slug: it.slug }), lastmod: it.updated_at }));
}