Missing kids selenium fixes

This commit is contained in:
Luciano Gervasoni
2025-07-08 09:43:40 +02:00
parent e81a96f4bd
commit 522c1cb8b3
8 changed files with 105 additions and 37 deletions

View File

@@ -57,15 +57,18 @@ class MissingKidsFetcher():
# Find all <img> tags with src attributes. Extract src URLs
image_urls = [img.get_attribute("src") for img in driver.find_elements(By.XPATH, "//img[@src]")]
# Redirects to 404?
if ("missingkids.org/404" in driver.current_url) or (any(["thumb-404.png" in i for i in image_urls])):
# Status invalid
results = {"status": "invalid"}
elif ("Have you seen this child?" in driver.title):
# Status valid
results = {"status": "valid"}
# Redirection to valid URL? -> Duplicate
elif (driver.current_url != url):
# Redirection (duplicate)
results = {"status": "duplicate", "redirection": driver.current_url}
# Valid
elif ("Have you seen this child?" in driver.title):
# Status valid
results = {"status": "valid"}
else:
results = {"status": "unknown"}
except Exception as e: