goodgo-platform/scripts/backfill-osm-provinces.ts

/**
 * One-shot backfill for OSM-imported `IndustrialPark` rows whose
 * `province` is "Chưa xác định" (the placeholder we wrote when the OSM
 * tags lacked any addr:* hints).
 *
 *   Usage:
 *     NODE_OPTIONS="-r dotenv/config" DOTENV_CONFIG_PATH=.env \
 *       pnpm tsx scripts/backfill-osm-provinces.ts [--dry-run]
 *
 *   What it does:
 *     1. Selects every row where dataSource = 'OSM' AND province =
 *        'Chưa xác định'.
 *     2. Reads the centroid via ST_X / ST_Y from the `location` Point.
 *     3. Looks up the nearest province from VN_PROVINCE_CENTROIDS.
 *     4. Updates the row in batches.
 *
 *   Safe to re-run: skips rows where province is already filled in.
 */
import 'dotenv/config';
import { PrismaPg } from '@prisma/adapter-pg';
import { PrismaClient } from '@prisma/client';
import pg from 'pg';
import { nearestProvince } from './data/vn-province-centroids';

const pool = new pg.Pool({ connectionString: process.env['DATABASE_URL'] });
const adapter = new PrismaPg(pool);
const prisma = new PrismaClient({ adapter });

const dryRun = process.argv.includes('--dry-run');

interface Row {
  id: string;
  lat: number;
  lng: number;
}

async function main(): Promise<void> {
  console.log('🔍 Finding OSM rows with province="Chưa xác định"…');

  const rows = await prisma.$queryRawUnsafe<Row[]>(
    `SELECT id, ST_Y(location::geometry) AS lat, ST_X(location::geometry) AS lng
     FROM "IndustrialPark"
     WHERE "dataSource"::text = 'OSM' AND province = 'Chưa xác định'`,
  );
  console.log(`  → ${rows.length} rows need a province.`);

  if (!rows.length) {
    console.log('✓ Nothing to do.');
    return;
  }

  const updates = new Map<string, string[]>();
  for (const row of rows) {
    const province = nearestProvince(row.lat, row.lng);
    if (!updates.has(province)) updates.set(province, []);
    updates.get(province)!.push(row.id);
  }

  // Sort by impact for the dry-run preview.
  const summary = Array.from(updates.entries()).sort((a, b) => b[1].length - a[1].length);
  console.log('  → Distribution by inferred province:');
  for (const [province, ids] of summary) {
    console.log(`     ${province.padEnd(24)} ${ids.length}`);
  }

  if (dryRun) {
    console.log('💡 --dry-run: no writes performed.');
    return;
  }

  let totalUpdated = 0;
  for (const [province, ids] of updates) {
    // UPDATE in batches of 500 ids to avoid huge IN-lists.
    for (let i = 0; i < ids.length; i += 500) {
      const batch = ids.slice(i, i + 500);
      const result = await prisma.industrialPark.updateMany({
        where: { id: { in: batch } },
        data: { province },
      });
      totalUpdated += result.count;
    }
  }
  console.log(`✓ Updated ${totalUpdated} rows.`);
}

main()
  .catch((err) => {
    console.error(err);
    process.exitCode = 1;
  })
  .finally(async () => {
    await prisma.$disconnect();
    await pool.end();
  });