Art-gallery/scripts/update-influences.js

require('dotenv').config();
const path = require('path');
const pool = require('../server/db');
const INFLUENCES = require('./art-influences-data');
const { savePaintingImages } = require('./image-fetcher');
const { discoverInfluencesForWork } = require('./influence-discovery');
const {
  normalizeInfluencedBy,
  loadMovements,
  resolveWorkRef,
  resolveAndInsertSource,
} = require('./influence-resolver');

const FETCH_IMAGES = process.argv.includes('--fetch-images');
const DISCOVER = process.argv.includes('--discover');
const DISCOVER_ONLY = process.argv.includes('--discover-only');
const LIMIT = parseInt(process.argv.find((a) => a.startsWith('--limit='))?.split('=')[1] || '0', 10);
const IMAGE_DIR = path.resolve(process.env.IMAGE_DIR || path.join(__dirname, '../data/images'));

async function applyEdge(edge, movementsByName, stats) {
  const work = await resolveWorkRef(
    pool,
    edge.work,
    movementsByName,
    FETCH_IMAGES,
    IMAGE_DIR,
    savePaintingImages
  );
  if (!work) {
    stats.failed += 1;
    console.warn(`✗ unresolved work: ${edge.work.artist} / ${edge.work.title}`);
    return;
  }

  const sources = normalizeInfluencedBy(edge.influencedBy);
  for (const sourceRef of sources) {
    const result = await resolveAndInsertSource(
      pool,
      work,
      sourceRef,
      edge,
      movementsByName,
      FETCH_IMAGES,
      IMAGE_DIR,
      savePaintingImages
    );
    if (!result.ok) {
      if (result.reason !== 'self' && result.reason !== 'self artist') {
        stats.failed += 1;
        console.warn(
          `✗ ${edge.work.artist} / ${edge.work.title} <- ${sourceRef.type || 'painting'} ${result.reason}`
        );
      } else {
        stats.skipped += 1;
      }
      continue;
    }
    if (result.inserted) {
      stats.added += 1;
      console.log(`→ ${edge.work.artist} / ${edge.work.title} ← [${sourceRef.type || 'painting'}] ${result.label}`);
    } else {
      stats.skipped += 1;
    }
  }

  if (DISCOVER || DISCOVER_ONLY) {
    const discovered = await discoverInfluencesForWork({
      artist: edge.work.artist,
      title: edge.work.title,
      year: edge.year ?? edge.work.year ?? work.year,
    });
    for (const sourceRef of discovered) {
      const result = await resolveAndInsertSource(
        pool,
        work,
        sourceRef,
        sourceRef,
        movementsByName,
        false,
        IMAGE_DIR,
        null
      );
      if (result.inserted) {
        stats.discovered += 1;
        console.log(`~ discovered ${edge.work.artist} / ${edge.work.title} ← [${sourceRef.type}] ${result.label}`);
      }
    }
  }
}

async function discoverCatalog(limit) {
  const { rows } = await pool.query(
    `SELECT p.id, p.title, p.year, a.name AS artist_name
     FROM paintings p
     JOIN artists a ON a.id = p.artist_id
     ORDER BY a.name, p.sort_order, p.year NULLS LAST`
  );

  const targets = limit > 0 ? rows.slice(0, limit) : rows;
  const movementsByName = await loadMovements(pool);
  let discovered = 0;

  for (const row of targets) {
    const refs = await discoverInfluencesForWork({
      artist: row.artist_name,
      title: row.title,
      year: row.year,
    });
    for (const sourceRef of refs) {
      const result = await resolveAndInsertSource(
        pool,
        { id: row.id, year: row.year, artist_id: null },
        sourceRef,
        sourceRef,
        movementsByName,
        false,
        IMAGE_DIR,
        null
      );
      if (result.inserted) {
        discovered += 1;
        console.log(`~ discovered ${row.artist_name} / ${row.title} ← [${sourceRef.type}] ${result.label}`);
      }
    }
  }

  return discovered;
}

async function main() {
  const tableCheck = await pool.query(`
    SELECT EXISTS (
      SELECT FROM information_schema.tables
      WHERE table_schema = 'public' AND table_name = 'painting_influence_sources'
    ) AS ok
  `);
  if (!tableCheck.rows[0]?.ok) {
    console.error('Run npm run dev:migrate:influence-sources before update-influences.');
    process.exit(1);
  }

  const movementsByName = await loadMovements(pool);
  const stats = { added: 0, skipped: 0, failed: 0, discovered: 0 };

  if (!DISCOVER_ONLY) {
    const edges = LIMIT > 0 ? INFLUENCES.slice(0, LIMIT) : INFLUENCES;
    for (const edge of edges) {
      await applyEdge(edge, movementsByName, stats);
    }
  }

  if (DISCOVER_ONLY) {
    stats.discovered += await discoverCatalog(LIMIT);
  }

  const [totalSources, totalLegacy, connected] = await Promise.all([
    pool.query('SELECT COUNT(*)::int AS n FROM painting_influence_sources'),
    pool.query('SELECT COUNT(*)::int AS n FROM painting_influences'),
    pool.query(`
      SELECT COUNT(DISTINCT a.id)::int AS n
      FROM artists a
      JOIN paintings p ON p.artist_id = a.id
      LEFT JOIN painting_influence_sources pis ON pis.painting_id = p.id
      LEFT JOIN painting_influences pi ON pi.painting_id = p.id OR pi.influenced_by_painting_id = p.id
      WHERE pis.id IS NOT NULL OR pi.id IS NOT NULL
    `),
  ]);

  console.log(
    `\nDone: ${stats.added} curated added, ${stats.discovered} discovered, ${stats.skipped} skipped, ${stats.failed} failed`
  );
  console.log(`Total influence sources: ${totalSources.rows[0].n} (legacy painting edges: ${totalLegacy.rows[0].n})`);
  console.log(`Artists connected to influence graph: ${connected.rows[0].n}`);

  await pool.end();
}

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