Remove dangling screenshot_enricher file. Moved to modules/screenshot_enricher

pull/189/head
Patrick Robertson 2025-02-04 18:16:03 +01:00
rodzic 91ca325fd5
commit 48abb5e66b
1 zmienionych plików z 0 dodań i 40 usunięć

Wyświetl plik

@ -1,40 +0,0 @@
from loguru import logger
import time, os
from selenium.common.exceptions import TimeoutException
from auto_archiver.core import Enricher
from ..utils import Webdriver, url as UrlUtil, random_str
from ..core import Media, Metadata
class ScreenshotEnricher(Enricher):
name = "screenshot_enricher"
@staticmethod
def configs() -> dict:
return {
"width": {"default": 1280, "help": "width of the screenshots"},
"height": {"default": 720, "help": "height of the screenshots"},
"timeout": {"default": 60, "help": "timeout for taking the screenshot"},
"sleep_before_screenshot": {"default": 4, "help": "seconds to wait for the pages to load before taking screenshot"},
"http_proxy": {"default": "", "help": "http proxy to use for the webdriver, eg http://proxy-user:password@proxy-ip:port"},
}
def enrich(self, to_enrich: Metadata) -> None:
url = to_enrich.get_url()
if UrlUtil.is_auth_wall(url):
logger.debug(f"[SKIP] SCREENSHOT since url is behind AUTH WALL: {url=}")
return
logger.debug(f"Enriching screenshot for {url=}")
with Webdriver(self.width, self.height, self.timeout, 'facebook.com' in url, http_proxy=self.http_proxy) as driver:
try:
driver.get(url)
time.sleep(int(self.sleep_before_screenshot))
screenshot_file = os.path.join(self.tmp_dir, f"screenshot_{random_str(8)}.png")
driver.save_screenshot(screenshot_file)
to_enrich.add_media(Media(filename=screenshot_file), id="screenshot")
except TimeoutException:
logger.info("TimeoutException loading page for screenshot")
except Exception as e:
logger.error(f"Got error while loading webdriver for screenshot enricher: {e}")