52 lines
1.8 KiB
TypeScript
52 lines
1.8 KiB
TypeScript
|
import { PowerbeijingCrawler } from './powerbeijing_target';
|
||
|
|
import * as puppeteer from 'puppeteer';
|
||
|
|
|
||
|
|
// Increase timeout to 60 seconds for network operations
|
||
|
|
jest.setTimeout(60000*5);
|
||
|
|
|
||
|
|
describe('PowerbeijingCrawler Real Site Test', () => {
|
||
|
|
let browser: puppeteer.Browser;
|
||
|
|
|
||
|
|
beforeAll(async () => {
|
||
|
|
browser = await puppeteer.launch({
|
||
|
|
headless: false, // Change to false to see browser UI
|
||
|
|
args: ['--no-sandbox', '--disable-setuid-sandbox'],
|
||
|
|
});
|
||
|
|
});
|
||
|
|
|
||
|
|
afterAll(async () => {
|
||
|
|
if (browser) {
|
||
|
|
await browser.close();
|
||
|
|
}
|
||
|
|
});
|
||
|
|
|
||
|
|
it('should visit website and list all found bid information', async () => {
|
||
|
|
console.log(`\nStarting crawl for: ${PowerbeijingCrawler.name}`);
|
||
|
|
console.log(`Target URL: ${PowerbeijingCrawler.url}`);
|
||
|
|
|
||
|
|
const results = await PowerbeijingCrawler.crawl(browser);
|
||
|
|
|
||
|
|
console.log(`\nSuccessfully found ${results.length} items:\n`);
|
||
|
|
console.log('----------------------------------------');
|
||
|
|
results.forEach((item, index) => {
|
||
|
|
console.log(`${index + 1}. [${item.publishDate.toLocaleDateString()}] ${item.title}`);
|
||
|
|
console.log(` Link: ${item.url}`);
|
||
|
|
console.log('----------------------------------------');
|
||
|
|
});
|
||
|
|
|
||
|
|
// Basic assertions to ensure crawler is working
|
||
|
|
expect(results).toBeDefined();
|
||
|
|
expect(Array.isArray(results)).toBeTruthy();
|
||
|
|
// Warn but don't fail if site returns 0 items (could be empty or changed structure)
|
||
|
|
if (results.length === 0) {
|
||
|
|
console.warn('Warning: No items found. Check if website structure has changed or if list is currently empty.');
|
||
|
|
} else {
|
||
|
|
// Check data integrity of first item
|
||
|
|
const firstItem = results[0];
|
||
|
|
expect(firstItem.title).toBeTruthy();
|
||
|
|
expect(firstItem.url).toMatch(/^https?:\/\//);
|
||
|
|
expect(firstItem.publishDate).toBeInstanceOf(Date);
|
||
|
|
}
|
||
|
|
});
|
||
|
|
});
|