goodgo-platform/scripts/prune-non-vietnam-osm.ts

/**
 * Prune `IndustrialPark` rows whose centroid is outside the Vietnam
 * mainland polygon. Catches the cross-border bleed (Laos, Thailand,
 * Cambodia) that the Overpass bbox sync inevitably picks up.
 *
 *   Usage:
 *     NODE_OPTIONS="-r dotenv/config" DOTENV_CONFIG_PATH=.env \
 *       pnpm tsx scripts/prune-non-vietnam-osm.ts [--dry-run]
 *
 *   Strategy:
 *     1. Build a PostGIS polygon from `VN_COUNTRY_POLYGON_GEOJSON`.
 *     2. SELECT rows where `NOT ST_Within(location, polygon)`, scoped to
 *        OSM-sourced rows (we never want to delete a manually-curated
 *        row even if its centroid is wonky).
 *     3. DELETE in one statement (cascade removes any IndustrialListing
 *        rows attached to those parks).
 *
 * Safe to re-run: idempotent.
 */
import 'dotenv/config';
import { PrismaPg } from '@prisma/adapter-pg';
import { PrismaClient } from '@prisma/client';
import pg from 'pg';
import { VN_COUNTRY_POLYGON_GEOJSON } from './data/vn-country-polygon';

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');

async function main(): Promise<void> {
  const polygonSql = `ST_SetSRID(ST_GeomFromGeoJSON('${VN_COUNTRY_POLYGON_GEOJSON.replace(
    /'/g,
    "''",
  )}'), 4326)`;

  const outsideRows = await prisma.$queryRawUnsafe<
    { id: string; name: string; province: string; lat: number; lng: number; ha: number }[]
  >(
    `SELECT id, name, province,
            ROUND(ST_Y(location::geometry)::numeric, 3)::float AS lat,
            ROUND(ST_X(location::geometry)::numeric, 3)::float AS lng,
            COALESCE("totalAreaHa", 0) AS ha
       FROM "IndustrialPark"
      WHERE "dataSource" IN ('OSM', 'OSM_PROMOTED')
        AND NOT ST_Within(location::geometry, ${polygonSql})
      ORDER BY ha DESC NULLS LAST`,
  );

  console.log(`📍 Found ${outsideRows.length} OSM rows OUTSIDE the VN polygon.`);

  if (outsideRows.length === 0) {
    console.log('✓ Catalog is clean.');
    return;
  }

  // Show the top 15 by area so the operator can sanity-check before deleting.
  console.log('  Top 15 by area (will be deleted):');
  for (const row of outsideRows.slice(0, 15)) {
    console.log(
      `    ${row.name.slice(0, 50).padEnd(50)} ${row.province.slice(0, 16).padEnd(16)} ${
        row.ha
      } ha   (${row.lat}, ${row.lng})`,
    );
  }

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

  console.log(`\n🗑  Deleting ${outsideRows.length} rows…`);
  const result = await prisma.$executeRawUnsafe(
    `DELETE FROM "IndustrialPark"
      WHERE "dataSource" IN ('OSM', 'OSM_PROMOTED')
        AND NOT ST_Within(location::geometry, ${polygonSql})`,
  );
  console.log(`✓ Removed ${result} rows.`);
}

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