_group.add_argument('--stream', action='store_true',
help="""
Store the file content and display the resulting manifest on
-stdout. Do not write the manifest to Keep or save a Collection object
-in Arvados.
+stdout. Do not save a Collection object in Arvados.
""")
_group.add_argument('--as-manifest', action='store_true', dest='manifest',
""")
_group = upload_opts.add_mutually_exclusive_group()
-_group.add_argument('--trash-at', metavar='YYYY-MM-DD HH:MM', default=None,
+_group.add_argument('--trash-at', metavar='YYYY-MM-DDTHH:MM', default=None,
help="""
Set the trash date of the resulting collection to an absolute date in the future.
-The accepted format is defined by the ISO 8601 standard.
+The accepted format is defined by the ISO 8601 standard. Examples: 20090103, 2009-01-03, 20090103T181505, 2009-01-03T18:15:05.\n
+Timezone information can be added. If not, the provided date/time is assumed as being in the local system's timezone.
""")
_group.add_argument('--trash-after', type=int, metavar='DAYS', default=None,
help="""
self.exclude_names = exclude_names
self._trash_at = trash_at
- if self._trash_at is not None and type(self._trash_at) not in [datetime.datetime, datetime.timedelta]:
- raise TypeError('trash_at should be None, datetime or timedelta')
+ if self._trash_at is not None:
+ if type(self._trash_at) not in [datetime.datetime, datetime.timedelta]:
+ raise TypeError('trash_at should be None, timezone-naive datetime or timedelta')
+ if type(self._trash_at) == datetime.datetime and self._trash_at.tzinfo is not None:
+ raise TypeError('provided trash_at datetime should be timezone-naive')
if not self.use_cache and self.resume:
raise ArvPutArgumentConflict('resume cannot be True when use_cache is False')
self._save_state()
except Exception as e:
self.logger.error("Unexpected error trying to save cache file: {}".format(e))
+ # Keep remote collection's trash_at attribute synced when using relative expire dates
+ if self._remote_collection is not None and type(self._trash_at) == datetime.timedelta:
+ try:
+ self._api_client.collections().update(
+ uuid=self._remote_collection.manifest_locator(),
+ body={'trash_at': self._collection_trash_at().strftime("%Y-%m-%dT%H:%M:%S.%fZ")}
+ ).execute(num_retries=self.num_retries)
+ except Exception as e:
+ self.logger.error("Unexpected error trying to update remote collection's expire date: {}".format(e))
else:
self.bytes_written = self.bytes_skipped
# Call the reporter, if any
def _write_stdin(self, filename):
output = self._local_collection.open(filename, 'wb')
- self._write(sys.stdin, output)
+ self._write(sys.stdin.buffer, output)
output.close()
def _check_file(self, source, filename):
update_collection):
try:
self._remote_collection = arvados.collection.Collection(
- update_collection, api_client=self._api_client)
+ update_collection,
+ api_client=self._api_client,
+ num_retries=self.num_retries)
except arvados.errors.ApiError as error:
raise CollectionUpdateError("Cannot read collection {} ({})".format(update_collection, error))
else:
self._state['manifest'],
replication_desired=self.replication_desired,
put_threads=self.put_threads,
- api_client=self._api_client)
+ api_client=self._api_client,
+ num_retries=self.num_retries)
def _cached_manifest_valid(self):
"""
def collection_name(self):
return self._my_collection().api_response()['name'] if self._my_collection().api_response() else None
+ def collection_trash_at(self):
+ return self._my_collection().get_trash_at()
+
def manifest_locator(self):
return self._my_collection().manifest_locator()
# Trash arguments validation
trash_at = None
if args.trash_at is not None:
+ # ciso8601 considers YYYYMM as invalid but YYYY-MM as valid, so here we
+ # make sure the user provides a complete YYYY-MM-DD date.
+ if not re.match(r'^\d{4}(?P<dash>-?)\d{2}?(?P=dash)\d{2}', args.trash_at):
+ logger.error("--trash-at argument format invalid, use --help to see examples.")
+ sys.exit(1)
+ # Check if no time information was provided. In that case, assume end-of-day.
+ if re.match(r'^\d{4}(?P<dash>-?)\d{2}?(?P=dash)\d{2}$', args.trash_at):
+ args.trash_at += 'T23:59:59'
try:
trash_at = ciso8601.parse_datetime(args.trash_at)
except:
- logger.error("--trash-at argument format invalid, should be YYYY-MM-DDTHH:MM.")
+ logger.error("--trash-at argument format invalid, use --help to see examples.")
sys.exit(1)
else:
if trash_at.tzinfo is not None:
- # Timezone-aware datetime provided, convert to non-aware UTC
- delta = trash_at.tzinfo.utcoffset(None)
- trash_at = trash_at.replace(tzinfo=None) - delta
+ # Timezone aware datetime provided.
+ utcoffset = -trash_at.utcoffset()
+ else:
+ # Timezone naive datetime provided. Assume is local.
+ if time.daylight:
+ utcoffset = datetime.timedelta(seconds=time.altzone)
+ else:
+ utcoffset = datetime.timedelta(seconds=time.timezone)
+ # Convert to UTC timezone naive datetime.
+ trash_at = trash_at.replace(tzinfo=None) + utcoffset
+
if trash_at <= datetime.datetime.utcnow():
- logger.error("--trash-at argument should be set in the future")
+ logger.error("--trash-at argument must be set in the future")
sys.exit(1)
if args.trash_after is not None:
if args.trash_after < 1:
- logger.error("--trash-after argument should be >= 1")
+ logger.error("--trash-after argument must be >= 1")
sys.exit(1)
trash_at = datetime.timedelta(seconds=(args.trash_after * 24 * 60 * 60))
output = ','.join(writer.data_locators())
else:
try:
+ expiration_notice = ""
+ if writer.collection_trash_at() is not None:
+ # Get the local timezone-naive version, and log it with timezone information.
+ if time.daylight:
+ local_trash_at = writer.collection_trash_at().replace(tzinfo=None) - datetime.timedelta(seconds=time.altzone)
+ else:
+ local_trash_at = writer.collection_trash_at().replace(tzinfo=None) - datetime.timedelta(seconds=time.timezone)
+ expiration_notice = ". It will expire on {} {}.".format(
+ local_trash_at.strftime("%Y-%m-%d %H:%M:%S"), time.strftime("%z"))
if args.update_collection:
- logger.info(u"Collection updated: '{}'".format(writer.collection_name()))
+ logger.info(u"Collection updated: '{}'{}".format(
+ writer.collection_name(), expiration_notice))
else:
- logger.info(u"Collection saved as '{}'".format(writer.collection_name()))
+ logger.info(u"Collection saved as '{}'{}".format(
+ writer.collection_name(), expiration_notice))
if args.portable_data_hash:
output = writer.portable_data_hash()
else: