mirror
/
auto-archiver
kopia lustrzana https://github.com/bellingcat/auto-archiver
1
0
Forkuj 0
Kod Zgłoszenia Packages Projekty Wydania Wiki Aktywność

minor improvements

pull/40/head
msramalho 2022-06-15 17:14:08 +02:00
rodzic 5cc21fa4e0
commit b1f70bb818
1 zmienionych plików z 4 dodań i 2 usunięć
Pokaż statystyki Ściągnij plik aktualizacji Ściągnij plik porównania Expand all files Collapse all files

6
archivers/wayback_archiver.py
Wyświetl plik

@ -18,10 +18,12 @@ class WaybackArchiver(Archiver):
def __init__(self, storage: Storage, driver, config: WaybackConfig): def __init__(self, storage: Storage, driver, config: WaybackConfig):
super(WaybackArchiver, self).__init__(storage, driver) super(WaybackArchiver, self).__init__(storage, driver)
self.config = config self.config = config
# TODO: this logic should live at the auto-archiver level
self.seen_urls = {} self.seen_urls = {}
def download(self, url, check_if_exists=False): def download(self, url, check_if_exists=False):
if self.config is None:
logger.error('Missing Wayback config')
return False
if check_if_exists: if check_if_exists:
if url in self.seen_urls: return self.seen_urls[url] if url in self.seen_urls: return self.seen_urls[url]
@ -57,7 +59,7 @@ class WaybackArchiver(Archiver):
retries += 1 retries += 1
if status_r.status_code != 200: if status_r.status_code != 200:
return ArchiveResult(status="Internet archive failed", screenshot=screenshot) return ArchiveResult(status=f"Internet archive failed: check https://web.archive.org/save/status/{job_id}", screenshot=screenshot)
status_json = status_r.json() status_json = status_r.json()
if status_json['status'] != 'success': if status_json['status'] != 'success':