ScraperControl/scripts/dedup-mass-schedules.ts

#!/usr/bin/env tsx
import dotenv from 'dotenv';
import path from 'path';
dotenv.config({ path: path.resolve(process.cwd(), '.env.local') });
import { Pool } from 'pg';

const pool = new Pool({ connectionString: process.env.DATABASE_URL });

interface CountResult {
  churches_with_dups: string;
  duplicate_rows: string;
}

async function main() {
  const dryRun = !process.argv.includes('--execute');

  if (dryRun) {
    console.log('DRY RUN - pass --execute to actually delete duplicates\n');
  }

  const client = await pool.connect();

  try {
    const countResult = await client.query<CountResult>(`
      WITH ranked AS (
        SELECT church_id,
               ROW_NUMBER() OVER (
                 PARTITION BY church_id, day_of_week, time, language
                 ORDER BY created_at ASC
               ) AS rn
        FROM mass_schedules
        WHERE is_active = true
      )
      SELECT COUNT(DISTINCT church_id) AS churches_with_dups,
             COUNT(*) AS duplicate_rows
      FROM ranked
      WHERE rn > 1;
    `);

    const { churches_with_dups, duplicate_rows } = countResult.rows[0];
    console.log(`Churches with duplicate schedules: ${churches_with_dups}`);
    console.log(`Duplicate rows to ${dryRun ? 'delete' : 'delete'}: ${duplicate_rows}\n`);

    if (!dryRun && Number(duplicate_rows) > 0) {
      console.log('Deleting duplicates (keeping oldest by created_at)...');

      const deleteResult = await client.query(`
        WITH ranked AS (
          SELECT id,
                 ROW_NUMBER() OVER (
                   PARTITION BY church_id, day_of_week, time, language
                   ORDER BY created_at ASC
                 ) AS rn
          FROM mass_schedules
          WHERE is_active = true
        )
        DELETE FROM mass_schedules
        WHERE id IN (SELECT id FROM ranked WHERE rn > 1);
      `);

      console.log(`Deleted ${deleteResult.rowCount} duplicate mass schedule rows.`);
    }
  } finally {
    client.release();
    await pool.end();
  }
}

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