def get_corpus_status(self):
'''
Return status data for the corpus of the specified S3 bucket/project
@type bucket_name: String
@param bucket_name: Name of the S3 bucket to use
@type project_name: String
@param project_name: Name of the project folder inside the S3 bucket
@rtype: dict
@return: Dictionary containing corpus size per date modified
'''
remote_keys = list(self.bucket.list(self.remote_path_corpus))
status_data = {}
for remote_key in remote_keys:
# Ignore any folders
if remote_key.name.endswith("/"):
continue
dt = boto_parse_ts(remote_key.last_modified)
date_str = "%s-%02d-%02d" % (dt.year, dt.month, dt.day)
if date_str not in status_data:
status_data[date_str] = 0
status_data[date_str] += 1
return status_data
def get_corpus_status(self):
'''
Return status data for the corpus of the specified S3 bucket/project
@type bucket_name: String
@param bucket_name: Name of the S3 bucket to use
@type project_name: String
@param project_name: Name of the project folder inside the S3 bucket
@rtype: dict
@return: Dictionary containing corpus size per date modified
'''
remote_keys = list(self.bucket.list(self.remote_path_corpus))
status_data = {}
for remote_key in remote_keys:
# Ignore any folders
if remote_key.name.endswith("/"):
continue
dt = boto_parse_ts(remote_key.last_modified)
date_str = "%s-%02d-%02d" % (dt.year, dt.month, dt.day)
if date_str not in status_data:
status_data[date_str] = 0
status_data[date_str] += 1
return status_data