cleanup wacz patch

pull/137/head v0.9.4
msramalho 2024-02-21 18:14:30 +00:00
rodzic 5bf0a0206d
commit 5324d562ba
3 zmienionych plików z 19 dodań i 9 usunięć

Wyświetl plik

@ -90,7 +90,9 @@ class ArchivingOrchestrator:
if cached_result:
logger.debug("Found previously archived entry")
for d in self.databases:
d.done(cached_result, cached=True)
try: d.done(cached_result, cached=True)
except Exception as e:
logger.error(f"ERROR database {d.name}: {e}: {traceback.format_exc()}")
return cached_result
# 3 - call archivers until one succeeds
@ -120,6 +122,9 @@ class ArchivingOrchestrator:
result.status = "nothing archived"
# signal completion to databases and archivers
for d in self.databases: d.done(result)
for d in self.databases:
try: d.done(result)
except Exception as e:
logger.error(f"ERROR database {d.name}: {e}: {traceback.format_exc()}")
return result

Wyświetl plik

@ -39,14 +39,19 @@ class WaczArchiverEnricher(Enricher, Archiver):
def setup(self) -> None:
self.use_docker = os.environ.get('WACZ_ENABLE_DOCKER') or not os.environ.get('RUNNING_IN_DOCKER')
self.browsertrix_home_host = os.environ.get('BROWSERTRIX_HOME_HOST')
self.browsertrix_home_container = os.environ.get('BROWSERTRIX_HOME_CONTAINER') or self.browsertrix_home_host
# create crawls folder if not exists, so it can be safely removed in cleanup
if self.use_docker and self.browsertrix_home_host:
os.makedirs(self.browsertrix_home_host, exist_ok=True)
if self.use_docker:
if self.browsertrix_home_container:
os.makedirs(self.browsertrix_home_container, exist_ok=True)
def cleanup(self) -> None:
if self.use_docker and self.browsertrix_home_host:
logger.debug(f"Removing {self.browsertrix_home_host=}")
shutil.rmtree(self.browsertrix_home_host, ignore_errors=True)
if self.use_docker:
if self.browsertrix_home_container:
logger.debug(f"Removing {self.browsertrix_home_container=}")
shutil.rmtree(self.browsertrix_home_container, ignore_errors=True)
def download(self, item: Metadata) -> Metadata:
# this new Metadata object is required to avoid duplication
@ -64,7 +69,7 @@ class WaczArchiverEnricher(Enricher, Archiver):
collection = random_str(8)
browsertrix_home_host = self.browsertrix_home_host or os.path.abspath(ArchivingContext.get_tmp_dir())
browsertrix_home_container = os.environ.get('BROWSERTRIX_HOME_CONTAINER') or browsertrix_home_host
browsertrix_home_container = self.browsertrix_home_container or browsertrix_home_host
cmd = [
"crawl",

Wyświetl plik

@ -3,7 +3,7 @@ _MAJOR = "0"
_MINOR = "9"
# On main and in a nightly release the patch should be one ahead of the last
# released build.
_PATCH = "3"
_PATCH = "4"
# This is mainly for nightly builds which have the suffix ".dev$DATE". See
# https://semver.org/#is-v123-a-semantic-version for the semantics.
_SUFFIX = ""