From 5dac180f486cbc6bb202debd5dde996a9c8204b4 Mon Sep 17 00:00:00 2001 From: MohamedBassem Date: Tue, 23 Apr 2024 21:22:53 +0100 Subject: fix(crawler): Better extraction for amazon images --- apps/workers/crawlerWorker.ts | 2 ++ apps/workers/package.json | 1 + 2 files changed, 3 insertions(+) (limited to 'apps/workers') diff --git a/apps/workers/crawlerWorker.ts b/apps/workers/crawlerWorker.ts index 2d1c7564..b583864d 100644 --- a/apps/workers/crawlerWorker.ts +++ b/apps/workers/crawlerWorker.ts @@ -10,6 +10,7 @@ import { eq } from "drizzle-orm"; import { isShuttingDown } from "exit"; import { JSDOM } from "jsdom"; import metascraper from "metascraper"; +import metascraperAmazon from "metascraper-amazon"; import metascraperDescription from "metascraper-description"; import metascraperImage from "metascraper-image"; import metascraperLogo from "metascraper-logo-favicon"; @@ -37,6 +38,7 @@ import { } from "@hoarder/shared/queues"; const metascraperParser = metascraper([ + metascraperAmazon(), metascraperReadability(), metascraperTitle(), metascraperDescription(), diff --git a/apps/workers/package.json b/apps/workers/package.json index e14c576b..7975cc84 100644 --- a/apps/workers/package.json +++ b/apps/workers/package.json @@ -16,6 +16,7 @@ "drizzle-orm": "^0.29.4", "jsdom": "^24.0.0", "metascraper": "^5.43.4", + "metascraper-amazon": "^5.45.0", "metascraper-description": "^5.43.4", "metascraper-image": "^5.43.4", "metascraper-logo": "^5.43.4", -- cgit v1.2.3-70-g09d2