aboutsummaryrefslogtreecommitdiffstats
path: root/packages/workers
diff options
context:
space:
mode:
authorMohamedBassem <me@mbassem.com>2024-03-05 00:33:59 +0000
committerMohamedBassem <me@mbassem.com>2024-03-05 00:33:59 +0000
commit591358c5a54919b7b8fc0819679c11ffe1d3472e (patch)
tree9095a4c8aba866f2c7d75408a20432920518b22e /packages/workers
parentfdd60a5764910619f9de6e8671767bafebaf3d66 (diff)
downloadkarakeep-591358c5a54919b7b8fc0819679c11ffe1d3472e.tar.zst
fix: Use puppeteer adblocker to block cookies notices
Diffstat (limited to 'packages/workers')
-rw-r--r--packages/workers/crawler.ts6
-rw-r--r--packages/workers/package.json1
2 files changed, 7 insertions, 0 deletions
diff --git a/packages/workers/crawler.ts b/packages/workers/crawler.ts
index f1ee07f3..fbbee730 100644
--- a/packages/workers/crawler.ts
+++ b/packages/workers/crawler.ts
@@ -18,6 +18,7 @@ import { db } from "@hoarder/db";
import { Browser } from "puppeteer";
import puppeteer from "puppeteer-extra";
import StealthPlugin from "puppeteer-extra-plugin-stealth";
+import AdblockerPlugin from "puppeteer-extra-plugin-adblocker";
import metascraper from "metascraper";
@@ -70,6 +71,11 @@ async function launchBrowser() {
export class CrawlerWorker {
static async build() {
puppeteer.use(StealthPlugin());
+ puppeteer.use(
+ AdblockerPlugin({
+ blockTrackersAndAnnoyances: true,
+ }),
+ );
await launchBrowser();
logger.info("Starting crawler worker ...");
diff --git a/packages/workers/package.json b/packages/workers/package.json
index a7b62462..f2fc164c 100644
--- a/packages/workers/package.json
+++ b/packages/workers/package.json
@@ -26,6 +26,7 @@
"openai": "^4.26.1",
"puppeteer": "^22.0.0",
"puppeteer-extra": "^3.3.6",
+ "puppeteer-extra-plugin-adblocker": "^2.13.6",
"puppeteer-extra-plugin-stealth": "^2.11.2",
"tsx": "^4.7.1",
"typescript": "^5",