ScraperControl/scripts/debug/test-french-scraper.ts

#!/usr/bin/env tsx
/**
 * Test international scraper against French churches
 */

import { config } from 'dotenv';
config({ path: '.env.local' });
config({ path: '.env' });

import { Pool } from 'pg';
import { PrismaPg } from '@prisma/adapter-pg';
import { PrismaClient } from '@prisma/client';
import { GenericScraper } from '../../src/scrapers/strategies/generic';

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

async function testFrenchScraper() {
  console.log('Testing French church mass schedule scraping...\n');

  // Get French churches with websites
  const churches = await prisma.church.findMany({
    where: {
      country: 'FR',
      website: { not: null },
      source: 'osm',
    },
    take: 5,
    orderBy: { createdAt: 'asc' },
  });

  if (churches.length === 0) {
    console.log('No French churches with websites found.');
    await prisma.$disconnect();
    await pool.end();
    return;
  }

  console.log(`Found ${churches.length} French churches to test:\n`);

  const scraper = new GenericScraper();
  await scraper.init();
  scraper.setCountry('FR');

  let successCount = 0;
  let failCount = 0;

  for (const church of churches) {
    console.log(`━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━`);
    console.log(`Church: ${church.name}`);
    console.log(`City: ${church.city || 'Unknown'}`);
    console.log(`URL: ${church.website}`);
    console.log('');

    try {
      const result = await scraper.scrape(church.website!);

      if (result.success && result.schedules.length > 0) {
        successCount++;
        console.log(`✅ SUCCESS - Found ${result.schedules.length} schedules\n`);

        // Group by day and show
        const byDay: Record<number, typeof result.schedules> = {};
        for (const sched of result.schedules) {
          if (!byDay[sched.dayOfWeek]) byDay[sched.dayOfWeek] = [];
          byDay[sched.dayOfWeek].push(sched);
        }

        const dayNames = ['Dimanche', 'Lundi', 'Mardi', 'Mercredi', 'Jeudi', 'Vendredi', 'Samedi'];
        Object.entries(byDay).forEach(([day, scheds]) => {
          console.log(`  ${dayNames[parseInt(day)]}:`);
          scheds.forEach(s => {
            console.log(`    ${s.time} - ${s.language || 'Unknown'} (${s.massType || 'Mass'})`);
          });
        });
        console.log('');
      } else {
        failCount++;
        console.log(`❌ FAILED - ${result.error}`);
        console.log('');
      }
    } catch (err: any) {
      failCount++;
      console.log(`❌ ERROR - ${err.message}`);
      console.log('');
    }
  }

  await scraper.close();

  console.log('━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━');
  console.log(`\nRESULTS: ${successCount}/${churches.length} successful (${((successCount / churches.length) * 100).toFixed(0)}%)`);
  console.log(`Success: ${successCount}, Failed: ${failCount}\n`);

  await prisma.$disconnect();
  await pool.end();
}

testFrenchScraper().catch(console.error);