bellingcat/auto-archiver
1
0
Fork 0
mirror of https://github.com/bellingcat/auto-archiver.git synced 2026-06-11 20:58:29 +03:00
Code Issues Packages Projects Releases Wiki Activity

Enricher tests

Browse Source
This commit is contained in:
erinhmclark
2025-02-12 19:32:40 +00:00
parent d9d936c2ca
commit cbe98c729d
7 changed files with 538 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/auto_archiver/modules/screenshot_enricher/screenshot_enricher.py
View File

@@ -11,6 +11,10 @@ from auto_archiver.core import Media, Metadata
class ScreenshotEnricher(Enricher):
def __init__(self, webdriver_factory=None):
super().__init__()
self.webdriver_factory = webdriver_factory or Webdriver
def enrich(self, to_enrich: Metadata) -> None:
url = to_enrich.get_url()
@@ -20,7 +24,8 @@ class ScreenshotEnricher(Enricher):
logger.debug(f"Enriching screenshot for {url=}")
auth = self.auth_for_site(url)
with Webdriver(self.width, self.height, self.timeout, facebook_accept_cookies='facebook.com' in url,
with self.webdriver_factory(
self.width, self.height, self.timeout, facebook_accept_cookies='facebook.com' in url,
http_proxy=self.http_proxy, print_options=self.print_options, auth=auth) as driver:
try:
driver.get(url)
@@ -38,3 +43,4 @@ class ScreenshotEnricher(Enricher):
logger.info("TimeoutException loading page for screenshot")
except Exception as e:
logger.error(f"Got error while loading webdriver for screenshot enricher: {e}")