vodorod/parcer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

chore: lint

Browse Source
This commit is contained in:
karishmas6
2024-08-20 23:15:33 +05:30
parent 6b6583647a
commit 6d5ffb5222
1 changed files with 34 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

46
maxun-core/src/interpret.ts
View File

@@ -393,40 +393,40 @@ export default class Interpreter extends EventEmitter {
break; break;
case 'scrollUp': case 'scrollUp':
break; break;
case 'clickNext': case 'clickNext':
const pageResults = await page.evaluate((cfg) => window.scrapeList(cfg), config); const pageResults = await page.evaluate((cfg) => window.scrapeList(cfg), config);
// Filter out already scraped items // Filter out already scraped items
const newResults = pageResults.filter(item => { const newResults = pageResults.filter(item => {
const uniqueKey = JSON.stringify(item); const uniqueKey = JSON.stringify(item);
if (scrapedItems.has(uniqueKey)) return false; // Ignore if already scraped if (scrapedItems.has(uniqueKey)) return false; // Ignore if already scraped
scrapedItems.add(uniqueKey); // Mark as scraped scrapedItems.add(uniqueKey); // Mark as scraped
return true; return true;
}); });
allResults = allResults.concat(newResults); allResults = allResults.concat(newResults);
if (config.limit && allResults.length >= config.limit) { if (config.limit && allResults.length >= config.limit) {
return allResults.slice(0, config.limit); return allResults.slice(0, config.limit);
} }
const nextButton = await page.$(config.pagination.selector); const nextButton = await page.$(config.pagination.selector);
if (!nextButton) { if (!nextButton) {
return allResults; // No more pages to scrape return allResults; // No more pages to scrape
} }
await Promise.all([ await Promise.all([
nextButton.click(), nextButton.click(),
page.waitForNavigation({ waitUntil: 'networkidle' }) page.waitForNavigation({ waitUntil: 'networkidle' })
]); ]);
await page.waitForTimeout(1000); await page.waitForTimeout(1000);
break; break;
case 'clickLoadMore': case 'clickLoadMore':
const loadMoreButton = await page.$(config.pagination.selector); const loadMoreButton = await page.$(config.pagination.selector);
if (!loadMoreButton) { if (!loadMoreButton) {