240e7f46f2
Routes requests through CT130/131/132 proxy pool (192.168.178.77/76/74:1080) when PROXY_URLS env var is set. Uses ProxyConfiguration from crawlee for PlaywrightCrawler scrapers and socks-proxy-agent for fetch-based scrapers.
33 lines
896 B
JSON
33 lines
896 B
JSON
{
|
|
"name": "@tip/scraper",
|
|
"version": "0.1.0",
|
|
"private": true,
|
|
"description": "TIP scraper engine — Crawlee + Playwright for competitor pricing, stock, datasheets, FAQs",
|
|
"main": "dist/index.js",
|
|
"scripts": {
|
|
"build": "tsc",
|
|
"dev": "tsx src/index.ts",
|
|
"scrape:fs": "tsx src/scrapers/fs-com.ts",
|
|
"scrape:cisco": "tsx src/scrapers/cisco-tmg.ts",
|
|
"scrape:optcore": "tsx src/scrapers/optcore.ts",
|
|
"scrape:news": "tsx src/scrapers/news.ts",
|
|
"scrape:all": "tsx src/index.ts --all"
|
|
},
|
|
"dependencies": {
|
|
"crawlee": "^3.12.0",
|
|
"playwright": "^1.50.0",
|
|
"pg": "^8.13.1",
|
|
"pg-boss": "^10.1.5",
|
|
"dotenv": "^16.4.7",
|
|
"cheerio": "^1.0.0",
|
|
"xml2js": "^0.6.2",
|
|
"socks-proxy-agent": "^8.0.5"
|
|
},
|
|
"devDependencies": {
|
|
"@types/pg": "^8.11.11",
|
|
"@types/xml2js": "^0.4.14",
|
|
"typescript": "^5.9.3",
|
|
"tsx": "^4.19.0"
|
|
}
|
|
}
|