42 lines
1.2 KiB
TypeScript
42 lines
1.2 KiB
TypeScript
|
#!/usr/bin/env tsx
|
||
|
|
import { GenericScraper } from '../../src/scrapers/strategies/generic';
|
||
|
|
|
||
|
|
async function main() {
|
||
|
|
const scraper = new GenericScraper();
|
||
|
|
await scraper.init();
|
||
|
|
scraper.setCountry('DE');
|
||
|
|
|
||
|
|
const result = await scraper.scrape('https://www.alterpeter.de/');
|
||
|
|
|
||
|
|
if (result.rawHtml) {
|
||
|
|
const text = result.rawHtml
|
||
|
|
.replace(/<script[^>]*>[\s\S]*?<\/script>/gi, '')
|
||
|
|
.replace(/<style[^>]*>[\s\S]*?<\/style>/gi, '')
|
||
|
|
.replace(/<[^>]+>/g, ' ')
|
||
|
|
.replace(/\s+/g, ' ')
|
||
|
|
.toLowerCase();
|
||
|
|
|
||
|
|
// Find all instances of "donnerstag" (Thursday)
|
||
|
|
let idx = 0;
|
||
|
|
let count = 0;
|
||
|
|
while ((idx = text.indexOf('donnerstag', idx)) !== -1) {
|
||
|
|
count++;
|
||
|
|
const contextBefore = text.substring(Math.max(0, idx - 100), idx);
|
||
|
|
const contextAfter = text.substring(idx, Math.min(text.length, idx + 200));
|
||
|
|
|
||
|
|
console.log(`=== Donnerstag occurrence ${count} at position ${idx} ===`);
|
||
|
|
console.log(`BEFORE: "...${contextBefore}"`);
|
||
|
|
console.log(`AFTER: "${contextAfter}..."`);
|
||
|
|
console.log('');
|
||
|
|
|
||
|
|
idx++;
|
||
|
|
}
|
||
|
|
|
||
|
|
console.log(`Total "donnerstag" occurrences: ${count}`);
|
||
|
|
}
|
||
|
|
|
||
|
|
await scraper.close();
|
||
|
|
}
|
||
|
|
|
||
|
|
main().catch(console.error);
|