github/v-firecrawl
github/v-firecrawl
0
Fork 0
Code Actions

[Feat] Added allowed urls

Browse Source 
FireCrawl should be able to scrape LinkedIn Articles (/pulse/*)
This commit is contained in:
rafaelsideguide 2024-04-25 08:39:45 -03:00
parent 75e82869ed
commit 75597f72a1
1 changed files with 8 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
apps/api/src/scraper/WebScraper/utils/blocklist.ts
View File

@ -14,6 +14,14 @@ const socialMediaBlocklist = [
'telegram.org', 'telegram.org',
]; ];
const allowedUrls = [
'linkedin.com/pulse'
];
export function isUrlBlocked(url: string): boolean { export function isUrlBlocked(url: string): boolean {
if (allowedUrls.some(allowedUrl => url.includes(allowedUrl))) {
return false;
}
return socialMediaBlocklist.some(domain => url.includes(domain)); return socialMediaBlocklist.some(domain => url.includes(domain));
} }