bellingcat/auto-archiver
1
0
Fork 0
mirror of https://github.com/bellingcat/auto-archiver.git synced 2026-06-13 05:38:29 +03:00
Code Issues Packages Projects Releases Wiki Activity

Provide better logs for screenshot enricher when auth is/isn't supported (cookies only)

Browse Source
This commit is contained in:
Patrick Robertson
2025-03-21 12:05:47 +04:00
parent 5b131996c6
commit 14c56f4916
3 changed files with 58 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
src/auto_archiver/modules/screenshot_enricher/screenshot_enricher.py
View File

@@ -19,12 +19,21 @@ class ScreenshotEnricher(Enricher):
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=}")
auth = self.auth_for_site(url)
# screenshot enricher only supports cookie-type auth (selenium)
has_valid_auth = auth and (auth.get("cookies") or auth.get("cookies_jar") or auth.get("cookie"))
if UrlUtil.is_auth_wall(url) and not has_valid_auth:
logger.warning(f"[SKIP] SCREENSHOT since url is behind AUTH WALL and no login details provided: {url=}")
if any(auth.get(key) for key in ["username", "password", "api_key", "api_secret"]):
logger.warning(
f"Screenshot enricher only supports cookie-type authentication, you have provided {auth.keys()} which are not supported.\
Consider adding 'cookie', 'cookies_file' or 'cookies_from_browser' to your auth for this site."
)
return
with self.webdriver_factory(
self.width,
self.height,