cannaiq/backend/src/scripts/run-backfill.ts

#!/usr/bin/env npx tsx
/**
 * Run Backfill CLI
 *
 * Import historical payloads from existing data sources.
 *
 * Usage:
 *   npx tsx src/scripts/run-backfill.ts [options]
 *
 * Options:
 *   --source SOURCE   Source to backfill from:
 *                     - dutchie_products (default)
 *                     - snapshots
 *                     - cache_files
 *                     - all
 *   --dry-run         Print changes without modifying DB
 *   --limit N         Max payloads to create (default: unlimited)
 *   --dispensary ID   Only backfill specific dispensary
 *   --cache-path PATH Path to cache files (default: ./cache/payloads)
 */

import { Pool } from 'pg';
import { runBackfill, BackfillOptions } from '../hydration';

async function main() {
  const args = process.argv.slice(2);

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

  let source: BackfillOptions['source'] = 'dutchie_products';
  const sourceIdx = args.indexOf('--source');
  if (sourceIdx !== -1 && args[sourceIdx + 1]) {
    source = args[sourceIdx + 1] as BackfillOptions['source'];
  }

  let limit: number | undefined;
  const limitIdx = args.indexOf('--limit');
  if (limitIdx !== -1 && args[limitIdx + 1]) {
    limit = parseInt(args[limitIdx + 1], 10);
  }

  let dispensaryId: number | undefined;
  const dispIdx = args.indexOf('--dispensary');
  if (dispIdx !== -1 && args[dispIdx + 1]) {
    dispensaryId = parseInt(args[dispIdx + 1], 10);
  }

  let cachePath: string | undefined;
  const cacheIdx = args.indexOf('--cache-path');
  if (cacheIdx !== -1 && args[cacheIdx + 1]) {
    cachePath = args[cacheIdx + 1];
  }

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

  try {
    console.log('='.repeat(60));
    console.log('BACKFILL RUNNER');
    console.log('='.repeat(60));
    console.log(`Source: ${source}`);
    console.log(`Dry run: ${dryRun}`);
    if (limit) console.log(`Limit: ${limit}`);
    if (dispensaryId) console.log(`Dispensary: ${dispensaryId}`);
    if (cachePath) console.log(`Cache path: ${cachePath}`);
    console.log('');

    const results = await runBackfill(pool, {
      dryRun,
      source,
      limit,
      dispensaryId,
      cachePath,
    });

    console.log('\nBackfill Results:');
    console.log('='.repeat(40));

    for (const result of results) {
      console.log(`\n${result.source}:`);
      console.log(`  Payloads created: ${result.payloadsCreated}`);
      console.log(`  Skipped: ${result.skipped}`);
      console.log(`  Errors: ${result.errors.length}`);
      console.log(`  Duration: ${result.durationMs}ms`);

      if (result.errors.length > 0) {
        console.log('  First 5 errors:');
        for (const err of result.errors.slice(0, 5)) {
          console.log(`    - ${err}`);
        }
      }
    }

    const totalCreated = results.reduce((sum, r) => sum + r.payloadsCreated, 0);
    console.log(`\nTotal payloads created: ${totalCreated}`);
  } catch (error: any) {
    console.error('Backfill error:', error.message);
    process.exit(1);
  } finally {
    await pool.end();
  }
}

main();