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

68
maxun-core/src/interpret.ts
View File

@@ -374,7 +374,7 @@ export default class Interpreter extends EventEmitter {
let previousHeight = 0; let previousHeight = 0;
let currentPage = 1; let currentPage = 1;
// track unique items per page to avoid re-scraping // track unique items per page to avoid re-scraping
let scrapedItems: Set<string> = new Set<string>(); let scrapedItems: Set<string> = new Set<string>();
while (true) { while (true) {
switch (config.pagination.type) { switch (config.pagination.type) {
@@ -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
const newResults = pageResults.filter(item => {
const uniqueKey = JSON.stringify(item);
if (scrapedItems.has(uniqueKey)) return false; // Ignore if already scraped
scrapedItems.add(uniqueKey); // Mark as scraped
return true;
});
allResults = allResults.concat(newResults);
if (config.limit && allResults.length >= config.limit) {
return allResults.slice(0, config.limit);
}
const nextButton = await page.$(config.pagination.selector);
if (!nextButton) {
return allResults; // No more pages to scrape
}
await Promise.all([
nextButton.click(),
page.waitForNavigation({ waitUntil: 'networkidle' })
]);
await page.waitForTimeout(1000);
break; // Filter out already scraped items
const newResults = pageResults.filter(item => {
const uniqueKey = JSON.stringify(item);
if (scrapedItems.has(uniqueKey)) return false; // Ignore if already scraped
scrapedItems.add(uniqueKey); // Mark as scraped
return true;
});
allResults = allResults.concat(newResults);
if (config.limit && allResults.length >= config.limit) {
return allResults.slice(0, config.limit);
}
const nextButton = await page.$(config.pagination.selector);
if (!nextButton) {
return allResults; // No more pages to scrape
}
await Promise.all([
nextButton.click(),
page.waitForNavigation({ waitUntil: 'networkidle' })
]);
await page.waitForTimeout(1000);
break;
case 'clickLoadMore': case 'clickLoadMore':
const loadMoreButton = await page.$(config.pagination.selector); const loadMoreButton = await page.$(config.pagination.selector);
if (!loadMoreButton) { if (!loadMoreButton) {