feat: initial commit
This commit is contained in:
30
service/scrapers/stealthy.py
Normal file
30
service/scrapers/stealthy.py
Normal file
@@ -0,0 +1,30 @@
|
||||
from __future__ import annotations
|
||||
|
||||
import asyncio
|
||||
import time
|
||||
|
||||
from scrapling import StealthyFetcher
|
||||
|
||||
from ..models.request import ScrapeRequest
|
||||
from ..models.response import ScrapeResponse
|
||||
from .base import BaseScraper
|
||||
|
||||
|
||||
class StealthyScraper(BaseScraper):
|
||||
"""Wraps Scrapling's StealthyFetcher — TLS fingerprint impersonation."""
|
||||
|
||||
async def scrape(self, req: ScrapeRequest) -> ScrapeResponse:
|
||||
start = time.perf_counter()
|
||||
fetcher = StealthyFetcher(auto_match=False)
|
||||
|
||||
kwargs: dict = {
|
||||
"url": req.url,
|
||||
"timeout": req.timeout / 1000,
|
||||
}
|
||||
if req.headers:
|
||||
kwargs["extra_headers"] = req.headers
|
||||
if req.proxy:
|
||||
kwargs["proxy"] = req.proxy
|
||||
|
||||
page = await asyncio.to_thread(fetcher.fetch, **kwargs)
|
||||
return self._build_response(req, page, "stealth", start)
|
||||
Reference in New Issue
Block a user