test-api.js

#!/usr/bin/env node

const axios = require('axios');

const API_BASE = 'http://localhost:3000/api';

async function testAPI() {
  console.log('🧪 Testing Advanced Crawler API\n');

  try {
    // Test 1: Health Check
    console.log('1. Testing health endpoint...');
    const health = await axios.get(`${API_BASE}/health`);
    console.log(`   ✅ Health: ${health.data.status}`);

    // Test 2: Crawler Test Endpoint
    console.log('2. Testing crawler test endpoint...');
    const crawlerTest = await axios.get(`${API_BASE}/crawler/test`);
    console.log(`   ✅ Crawler test: ${crawlerTest.data.message}`);

    // Test 3: Single URL Crawl
    console.log('3. Testing single URL crawl...');
    const crawlResult = await axios.post(`${API_BASE}/crawler/crawl`, {
      url: 'https://httpbin.org/html',
      timeout: 15000,
      extractHeadings: true,
      extractLinks: true
    });
    console.log(`   ✅ Crawl successful: ${crawlResult.data.url} (${crawlResult.data.status})`);
    console.log(`   📊 Duration: ${crawlResult.data.duration}ms`);
    console.log(`   📄 Title: ${crawlResult.data.html?.title || 'N/A'}`);

    // Test 4: Create Session
    console.log('4. Testing session creation...');
    const session = await axios.post(`${API_BASE}/crawler/session`, {
      userAgent: 'Test Bot 1.0',
      description: 'API test session'
    });
    console.log(`   ✅ Session created: ${session.data.sessionId}`);

    // Test 5: Get Session Info
    console.log('5. Testing session retrieval...');
    const sessionInfo = await axios.get(`${API_BASE}/crawler/session/${session.data.sessionId}`);
    console.log(`   ✅ Session info retrieved: ${sessionInfo.data.session.requestCount} requests`);

    // Test 6: Batch Crawl
    console.log('6. Testing batch crawl...');
    const batchResult = await axios.post(`${API_BASE}/crawler/batch`, {
      urls: [
        'https://httpbin.org/html',
        'https://httpbin.org/json'
      ],
      sessionId: session.data.sessionId,
      concurrency: 2,
      delay: 500
    });
    console.log(`   ✅ Batch crawl: ${batchResult.data.summary.successful}/${batchResult.data.summary.total} successful`);

    // Test 7: Service Status
    console.log('7. Testing service status...');
    const status = await axios.get(`${API_BASE}/crawler/status`);
    console.log(`   ✅ Service status: ${status.data.status}`);
    console.log(`   📈 Active sessions: ${status.data.statistics.activeSessions}`);

    // Test 8: Clean up session
    console.log('8. Cleaning up session...');
    await axios.delete(`${API_BASE}/crawler/session/${session.data.sessionId}`);
    console.log(`   ✅ Session deleted successfully`);

    console.log('\n🎉 All tests passed! The API is working correctly.');

  } catch (error) {
    console.error('\n❌ Test failed:', {
      message: error.message,
      status: error.response?.status,
      data: error.response?.data
    });
    process.exit(1);
  }
}

// Run tests if this file is executed directly
if (require.main === module) {
  testAPI();
}

module.exports = testAPI;