Art-gallery/scripts/expand-paintings.js

require('dotenv').config();
const path = require('path');
const pool = require('../server/db');
const ADDITIONS = require('./famous-paintings-data');
const { savePaintingImages } = require('./image-fetcher');

const MIN_PAINTINGS = parseInt(process.env.MIN_PAINTINGS || '6', 10);
const FETCH_IMAGES = process.argv.includes('--fetch-images');
const IMAGE_DIR = path.resolve(process.env.IMAGE_DIR || path.join(__dirname, '../data/images'));

function normalizeTitle(s) {
  return s.toLowerCase().replace(/[^a-z0-9]+/g, ' ').trim();
}

function titleMatches(existingNorm, title) {
  const norm = normalizeTitle(title);
  if (existingNorm.has(norm)) return true;
  for (const e of existingNorm) {
    if (e.includes(norm) || norm.includes(e)) return true;
  }
  return false;
}

async function main() {
  const artistsRes = await pool.query('SELECT id, name FROM artists ORDER BY name');
  const artistByName = new Map(artistsRes.rows.map((a) => [a.name, a]));

  const unknown = new Set();
  for (const entry of ADDITIONS) {
    if (!artistByName.has(entry.artist)) unknown.add(entry.artist);
  }
  if (unknown.size) {
    console.warn('Unknown artist names in data (skipped):', [...unknown].join(', '));
  }

  let inserted = 0;
  let skipped = 0;
  let imagesFetched = 0;
  const newIds = [];

  for (const artist of artistsRes.rows) {
    const countRes = await pool.query(
      'SELECT COUNT(*)::int AS n FROM paintings WHERE artist_id = $1',
      [artist.id]
    );
    let count = countRes.rows[0].n;
    if (count >= MIN_PAINTINGS) continue;

    const existingRes = await pool.query(
      'SELECT title FROM paintings WHERE artist_id = $1',
      [artist.id]
    );
    const existingNorm = new Set(existingRes.rows.map((r) => normalizeTitle(r.title)));

    const candidates = ADDITIONS.filter((e) => e.artist === artist.name);
    for (const entry of candidates) {
      if (count >= MIN_PAINTINGS) break;
      if (titleMatches(existingNorm, entry.title)) {
        skipped += 1;
        continue;
      }

      const wikiTitle = entry.wikipedia_title || entry.title;
      const sortRes = await pool.query(
        'SELECT COALESCE(MAX(sort_order), 0) + 1 AS next FROM paintings WHERE artist_id = $1',
        [artist.id]
      );
      const sortOrder = sortRes.rows[0].next;

      const insert = await pool.query(
        `INSERT INTO paintings (artist_id, title, year, wikipedia_title, sort_order)
         VALUES ($1, $2, $3, $4, $5)
         RETURNING id`,
        [artist.id, entry.title, entry.year ?? null, wikiTitle, sortOrder]
      );

      const paintingId = insert.rows[0].id;
      existingNorm.add(normalizeTitle(entry.title));
      count += 1;
      inserted += 1;
      newIds.push({ id: paintingId, artist: artist.name, title: entry.title, wikiTitle });

      console.log(`+ ${artist.name}: ${entry.title}${entry.year ? ` (${entry.year})` : ''}`);
    }
  }

  console.log(`\nInserted ${inserted} paintings (${skipped} duplicates skipped)`);

  if (FETCH_IMAGES && newIds.length) {
    console.log(`\nFetching images for ${newIds.length} new paintings…`);
    for (const row of newIds) {
      try {
        const base = `${row.artist}_${row.title}`.replace(/[^a-zA-Z0-9_-]/g, '_');
        const saved = await savePaintingImages(row.wikiTitle, base, IMAGE_DIR, {
          artistName: row.artist,
          paintingTitle: row.title,
        });
        if (saved.imagePath || saved.thumbnailPath) {
          await pool.query(
            `UPDATE paintings SET image_path = $1, thumbnail_path = $2 WHERE id = $3`,
            [saved.imagePath, saved.thumbnailPath, row.id]
          );
          imagesFetched += 1;
          console.log(`  ✓ image: ${row.artist} — ${row.title}`);
        } else {
          console.warn(`  ✗ no image: ${row.artist} — ${row.title}`);
        }
      } catch (err) {
        console.warn(`  ✗ ${row.artist} — ${row.title}: ${err.message}`);
      }
    }
    console.log(`\nImages fetched: ${imagesFetched}/${newIds.length}`);
  } else if (newIds.length) {
    console.log('Run with --fetch-images to download artwork files.');
  }

  const summary = await pool.query(`
    SELECT
      COUNT(*) FILTER (WHERE cnt < $1)::int AS below_min,
      COUNT(*) FILTER (WHERE cnt >= $1)::int AS at_min,
      MAX(cnt)::int AS max_paintings,
      ROUND(AVG(cnt), 1) AS avg_paintings
    FROM (
      SELECT COUNT(p.id)::int AS cnt
      FROM artists a
      LEFT JOIN paintings p ON p.artist_id = a.id
      GROUP BY a.id
    ) t
  `, [MIN_PAINTINGS]);
  console.log('\nCatalog summary:', summary.rows[0]);

  await pool.end();
}

main().catch((err) => {
  console.error(err);
  process.exit(1);
});