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

#!/usr/bin/env npx tsx
/**
 * Dutchie Discovery CLI
 *
 * Command-line interface for running the Dutchie store discovery pipeline.
 *
 * Usage:
 *   npx tsx src/scripts/run-discovery.ts <command> [options]
 *
 * Commands:
 *   discover:state <state>     - Discover all stores in a state (e.g., AZ)
 *   discover:city <city>       - Discover stores in a single city
 *   discover:full              - Run full discovery pipeline
 *   seed:cities <state>        - Seed known cities for a state
 *   stats                      - Show discovery statistics
 *   list                       - List discovered locations
 *
 * Examples:
 *   npx tsx src/scripts/run-discovery.ts discover:state AZ
 *   npx tsx src/scripts/run-discovery.ts discover:city phoenix --state AZ
 *   npx tsx src/scripts/run-discovery.ts seed:cities AZ
 *   npx tsx src/scripts/run-discovery.ts stats
 *   npx tsx src/scripts/run-discovery.ts list --status discovered --state AZ
 */

import { Pool } from 'pg';
import {
  runFullDiscovery,
  discoverCity,
  discoverState,
  getDiscoveryStats,
  seedKnownCities,
} from '../discovery';
import { getCitiesForState } from '../discovery/location-discovery';

// Parse command line arguments
function parseArgs() {
  const args = process.argv.slice(2);
  const command = args[0] || 'help';
  const positional: string[] = [];
  const flags: Record<string, string | boolean> = {};

  for (let i = 1; i < args.length; i++) {
    const arg = args[i];
    if (arg.startsWith('--')) {
      const [key, value] = arg.slice(2).split('=');
      if (value !== undefined) {
        flags[key] = value;
      } else if (args[i + 1] && !args[i + 1].startsWith('--')) {
        flags[key] = args[i + 1];
        i++;
      } else {
        flags[key] = true;
      }
    } else {
      positional.push(arg);
    }
  }

  return { command, positional, flags };
}

// Create database pool
function createPool(): Pool {
  const connectionString = process.env.DATABASE_URL;
  if (!connectionString) {
    console.error('ERROR: DATABASE_URL environment variable is required');
    process.exit(1);
  }
  return new Pool({ connectionString });
}

// Print help
function printHelp() {
  console.log(`
Dutchie Discovery CLI

Usage:
  npx tsx src/scripts/run-discovery.ts <command> [options]

Commands:
  discover:state <state>     Discover all stores in a state (e.g., AZ)
  discover:city <city>       Discover stores in a single city
  discover:full              Run full discovery pipeline
  seed:cities <state>        Seed known cities for a state
  stats                      Show discovery statistics
  list                       List discovered locations

Options:
  --state <code>             State code (e.g., AZ, CA, ON)
  --country <code>           Country code (default: US)
  --status <status>          Filter by status (discovered, verified, rejected, merged)
  --limit <n>                Limit results (default: varies by command)
  --dry-run                  Don't make any changes, just show what would happen
  --verbose                  Show detailed output

Examples:
  npx tsx src/scripts/run-discovery.ts discover:state AZ
  npx tsx src/scripts/run-discovery.ts discover:city phoenix --state AZ
  npx tsx src/scripts/run-discovery.ts seed:cities AZ
  npx tsx src/scripts/run-discovery.ts stats
  npx tsx src/scripts/run-discovery.ts list --status discovered --state AZ --limit 20
`);
}

// Main
async function main() {
  const { command, positional, flags } = parseArgs();

  if (command === 'help' || flags.help) {
    printHelp();
    process.exit(0);
  }

  const pool = createPool();

  try {
    switch (command) {
      case 'discover:state': {
        const stateCode = positional[0] || (flags.state as string);
        if (!stateCode) {
          console.error('ERROR: State code is required');
          console.error('Usage: discover:state <state>');
          process.exit(1);
        }

        console.log(`\nDiscovering stores in ${stateCode}...\n`);
        const result = await discoverState(pool, stateCode.toUpperCase(), {
          dryRun: Boolean(flags['dry-run']),
          verbose: Boolean(flags.verbose),
          cityLimit: flags.limit ? parseInt(flags.limit as string, 10) : 100,
        });

        console.log('\n=== DISCOVERY RESULTS ===');
        console.log(`Cities crawled: ${result.locations.length}`);
        console.log(`Locations found: ${result.totalLocationsFound}`);
        console.log(`Locations upserted: ${result.totalLocationsUpserted}`);
        console.log(`Duration: ${(result.durationMs / 1000).toFixed(1)}s`);
        break;
      }

      case 'discover:city': {
        const citySlug = positional[0];
        if (!citySlug) {
          console.error('ERROR: City slug is required');
          console.error('Usage: discover:city <city-slug> [--state AZ]');
          process.exit(1);
        }

        console.log(`\nDiscovering stores in ${citySlug}...\n`);
        const result = await discoverCity(pool, citySlug, {
          stateCode: flags.state as string,
          countryCode: (flags.country as string) || 'US',
          dryRun: Boolean(flags['dry-run']),
          verbose: Boolean(flags.verbose),
        });

        if (!result) {
          console.error(`City not found: ${citySlug}`);
          process.exit(1);
        }

        console.log('\n=== DISCOVERY RESULTS ===');
        console.log(`City: ${result.citySlug}`);
        console.log(`Locations found: ${result.locationsFound}`);
        console.log(`Locations upserted: ${result.locationsUpserted}`);
        console.log(`New: ${result.locationsNew}, Updated: ${result.locationsUpdated}`);
        console.log(`Duration: ${(result.durationMs / 1000).toFixed(1)}s`);
        if (result.errors.length > 0) {
          console.log(`Errors: ${result.errors.length}`);
          result.errors.forEach((e) => console.log(`  - ${e}`));
        }
        break;
      }

      case 'discover:full': {
        console.log('\nRunning full discovery pipeline...\n');
        const result = await runFullDiscovery(pool, {
          stateCode: flags.state as string,
          countryCode: (flags.country as string) || 'US',
          cityLimit: flags.limit ? parseInt(flags.limit as string, 10) : 50,
          skipCityDiscovery: Boolean(flags['skip-cities']),
          onlyStale: !flags.all,
          staleDays: flags['stale-days'] ? parseInt(flags['stale-days'] as string, 10) : 7,
          dryRun: Boolean(flags['dry-run']),
          verbose: Boolean(flags.verbose),
        });

        console.log('\n=== FULL DISCOVERY RESULTS ===');
        console.log(`Cities discovered: ${result.cities.citiesFound}`);
        console.log(`Cities upserted: ${result.cities.citiesUpserted}`);
        console.log(`Cities crawled: ${result.locations.length}`);
        console.log(`Total locations found: ${result.totalLocationsFound}`);
        console.log(`Total locations upserted: ${result.totalLocationsUpserted}`);
        console.log(`Duration: ${(result.durationMs / 1000).toFixed(1)}s`);
        break;
      }

      case 'seed:cities': {
        const stateCode = positional[0] || (flags.state as string);
        if (!stateCode) {
          console.error('ERROR: State code is required');
          console.error('Usage: seed:cities <state>');
          process.exit(1);
        }

        // Dynamically fetch cities from Dutchie
        console.log(`\nFetching cities for ${stateCode} from Dutchie...\n`);
        const cityNames = await getCitiesForState(stateCode.toUpperCase());

        if (cityNames.length === 0) {
          console.error(`No cities found for state: ${stateCode}`);
          process.exit(1);
        }

        const cities = cityNames.map(name => ({
          name,
          slug: name.toLowerCase().replace(/\s+/g, '-').replace(/[^a-z0-9-]/g, ''),
          stateCode: stateCode.toUpperCase(),
        }));

        console.log(`Seeding ${cities.length} cities for ${stateCode}...\n`);
        const result = await seedKnownCities(pool, cities);
        console.log(`Created: ${result.created} new cities`);
        console.log(`Updated: ${result.updated} existing cities`);
        break;
      }

      case 'stats': {
        console.log('\nFetching discovery statistics...\n');
        const stats = await getDiscoveryStats(pool);

        console.log('=== CITIES ===');
        console.log(`Total: ${stats.cities.total}`);
        console.log(`Crawled (24h): ${stats.cities.crawledLast24h}`);
        console.log(`Never crawled: ${stats.cities.neverCrawled}`);
        console.log('');
        console.log('=== LOCATIONS ===');
        console.log(`Total active: ${stats.locations.total}`);
        console.log(`Discovered: ${stats.locations.discovered}`);
        console.log(`Verified: ${stats.locations.verified}`);
        console.log(`Merged: ${stats.locations.merged}`);
        console.log(`Rejected: ${stats.locations.rejected}`);
        console.log('');
        console.log('=== BY STATE ===');
        stats.locations.byState.forEach((s) => {
          console.log(`  ${s.stateCode}: ${s.count}`);
        });
        break;
      }

      case 'list': {
        const status = flags.status as string;
        const stateCode = flags.state as string;
        const limit = flags.limit ? parseInt(flags.limit as string, 10) : 50;

        let whereClause = 'WHERE active = TRUE';
        const params: any[] = [];
        let paramIndex = 1;

        if (status) {
          whereClause += ` AND status = $${paramIndex}`;
          params.push(status);
          paramIndex++;
        }

        if (stateCode) {
          whereClause += ` AND state_code = $${paramIndex}`;
          params.push(stateCode.toUpperCase());
          paramIndex++;
        }

        params.push(limit);

        const { rows } = await pool.query(
          `
          SELECT id, platform, name, city, state_code, status, platform_menu_url, first_seen_at
          FROM dutchie_discovery_locations
          ${whereClause}
          ORDER BY first_seen_at DESC
          LIMIT $${paramIndex}
          `,
          params
        );

        console.log(`\nFound ${rows.length} locations:\n`);
        console.log('ID\tStatus\t\tState\tCity\t\tName');
        console.log('-'.repeat(80));
        rows.forEach((row: any) => {
          const cityDisplay = (row.city || '').substring(0, 12).padEnd(12);
          const nameDisplay = (row.name || '').substring(0, 30);
          console.log(
            `${row.id}\t${row.status.padEnd(12)}\t${row.state_code || 'N/A'}\t${cityDisplay}\t${nameDisplay}`
          );
        });
        break;
      }

      default:
        console.error(`Unknown command: ${command}`);
        printHelp();
        process.exit(1);
    }
  } catch (error: any) {
    console.error('ERROR:', error.message);
    if (flags.verbose) {
      console.error(error.stack);
    }
    process.exit(1);
  } finally {
    await pool.end();
  }
}

main();