ScraperControl/scripts/debug/list-church-websites.ts

import { config } from 'dotenv';
import { PrismaClient } from '@prisma/client';
import { Pool } from 'pg';
import { PrismaPg } from '@prisma/adapter-pg';

// Load .env.local first, then .env
config({ path: '.env.local' });
config({ path: '.env' });

const connectionString = process.env.DATABASE_URL;

if (!connectionString) {
  throw new Error('DATABASE_URL environment variable is not set');
}

const pool = new Pool({ connectionString });
const adapter = new PrismaPg(pool);
const prisma = new PrismaClient({ adapter });

async function listChurchWebsites() {
  try {
    console.log('Fetching churches from database...\n');

    const churches = await prisma.church.findMany({
      select: {
        id: true,
        name: true,
        city: true,
        state: true,
        country: true,
        website: true,
        googlePlaceId: true,
      },
      orderBy: [
        { country: 'asc' },
        { state: 'asc' },
        { city: 'asc' },
      ],
    });

    console.log(`Total churches: ${churches.length}`);

    const withWebsite = churches.filter(c => c.website);
    const withGoogle = churches.filter(c => c.googlePlaceId);
    const withoutWebsite = churches.filter(c => !c.website);

    console.log(`Churches with website: ${withWebsite.length}`);
    console.log(`Churches with Google Place ID: ${withGoogle.length}`);
    console.log(`Churches without website: ${withoutWebsite.length}\n`);

    // Group by country
    const byCountry = churches.reduce((acc, church) => {
      const country = church.country || 'Unknown';
      if (!acc[country]) {
        acc[country] = [];
      }
      acc[country].push(church);
      return acc;
    }, {} as Record<string, typeof churches>);

    // Write to file
    let output = '# Church Websites\n\n';
    output += `Generated: ${new Date().toISOString()}\n\n`;
    output += `## Summary\n`;
    output += `- Total churches: ${churches.length}\n`;
    output += `- With website: ${withWebsite.length} (${((withWebsite.length / churches.length) * 100).toFixed(1)}%)\n`;
    output += `- With Google Place ID: ${withGoogle.length} (${((withGoogle.length / churches.length) * 100).toFixed(1)}%)\n`;
    output += `- Without website: ${withoutWebsite.length} (${((withoutWebsite.length / churches.length) * 100).toFixed(1)}%)\n\n`;

    // Add country breakdown
    output += `## By Country\n\n`;
    Object.entries(byCountry)
      .sort(([, a], [, b]) => b.length - a.length)
      .forEach(([country, countryChurches]) => {
        const withSite = countryChurches.filter(c => c.website).length;
        const withGoogle = countryChurches.filter(c => c.googlePlaceId).length;
        output += `### ${country} (${countryChurches.length} churches)\n`;
        output += `- With website: ${withSite} (${((withSite / countryChurches.length) * 100).toFixed(1)}%)\n`;
        output += `- With Google Place ID: ${withGoogle} (${((withGoogle / countryChurches.length) * 100).toFixed(1)}%)\n\n`;
      });

    // List all websites
    output += `## All Websites\n\n`;
    Object.entries(byCountry)
      .sort(([a], [b]) => a.localeCompare(b))
      .forEach(([country, countryChurches]) => {
        output += `### ${country}\n\n`;
        countryChurches.forEach(church => {
          const location = [church.city, church.state, church.country].filter(Boolean).join(', ');
          if (church.website) {
            output += `- **${church.name}** (${location})\n`;
            output += `  - Website: ${church.website}\n`;
            if (church.googlePlaceId) {
              output += `  - Google Place ID: ${church.googlePlaceId}\n`;
            }
            output += `  - DB ID: ${church.id}\n\n`;
          }
        });
      });

    // List churches without websites
    output += `## Churches Without Websites\n\n`;
    Object.entries(byCountry)
      .sort(([a], [b]) => a.localeCompare(b))
      .forEach(([country, countryChurches]) => {
        const without = countryChurches.filter(c => !c.website);
        if (without.length > 0) {
          output += `### ${country}\n\n`;
          without.forEach(church => {
            const location = [church.city, church.state, church.country].filter(Boolean).join(', ');
            output += `- **${church.name}** (${location})\n`;
            if (church.googlePlaceId) {
              output += `  - Google Place ID: ${church.googlePlaceId}\n`;
            }
            output += `  - DB ID: ${church.id}\n\n`;
          });
        }
      });

    // Write to file
    const fs = await import('fs/promises');
    await fs.writeFile('church-websites.md', output);
    console.log('✓ Written to church-websites.md');

  } catch (error) {
    console.error('Error:', error);
    process.exit(1);
  } finally {
    await prisma.$disconnect();
    await pool.end();
  }
}

listChurchWebsites();