`scripts/ingest/region-images-fallback.js`

49 lines
'use strict';
// For regions without an image, fall back to the province image.
const { fetchJson, sleep, pool } = require('./_lib');

const PROVINCE_TITLES = {
  'San José':    'San_José_Province',
  'Alajuela':    'Alajuela_Province',
  'Heredia':     'Heredia_Province',
  'Cartago':     'Cartago_Province',
  'Limón':       'Limón_Province',
  'Puntarenas':  'Puntarenas_Province',
  'Guanacaste':  'Guanacaste_Province',
};

(async () => {
  const cache = {};
  for (const [prov, title] of Object.entries(PROVINCE_TITLES)) {
    try {
      const j = await fetchJson(`https://en.wikipedia.org/api/rest_v1/page/summary/${encodeURIComponent(title)}`,
        { headers: { 'Api-User-Agent': 'CR-Directory/0.1 (info@agentabrams.com)' } });
      const img = (j.originalimage || j.thumbnail);
      if (img) cache[prov] = {
        url: img.source,
        credit: `Wikipedia: ${j.title}`,
        source_url: j.content_urls?.desktop?.page || `https://en.wikipedia.org/wiki/${encodeURIComponent(title)}`,
      };
      console.log(`[prov] ${prov} → ${img ? 'HIT' : 'miss'}`);
      await sleep(400);
    } catch (e) { console.warn(`[prov] ${prov} err: ${e.message}`); }
  }

  const { rows } = await pool.query(
    `SELECT id, slug, name, province FROM regions WHERE region_type IN ('city','town','canton') AND image_url IS NULL`
  );
  let n = 0;
  for (const r of rows) {
    const fallback = cache[r.province];
    if (!fallback) continue;
    await pool.query(
      `UPDATE regions SET image_url=$1, image_credit=$2, image_license='CC via Wikimedia Commons',
              image_source_url=$3, image_fetched_at=NOW() WHERE id=$4`,
      [fallback.url, fallback.credit + ' (province fallback)', fallback.source_url, r.id]
    );
    n++;
  }
  console.log(`[fallback] applied province image to ${n} regions`);
  await pool.end();
})();