def dt():
jobs = []
if project.verbose: print("DT To BigQuery")
# legacy deprecated ( do not use )
if 'path' in project.task: project.task['paths'] = [project.task['path']]
# loop all dt files to match pattern or match any pattern
print('PATHS', project.task['paths'])
for path in (project.task['paths'] or ['']):
print(path)
for dt_object in object_list(project.task['auth'], '%s:%s' % (project.task['bucket'], path), raw=True):
dt_size = dt_object['size']
dt_file = dt_object['name']
dt_time = dt_timestamp(dt_file)
dt_partition = dt_file.split('.', 1)[0]
if ((project.task.get('days') is None and project.task.get('hours') is None) or (dt_time > project.now - timedelta(days=project.task.get('days', 60), hours=project.task.get('hours', 0)))):
if not table_exists(project.task['to']['auth'], project.id, project.task['to']['dataset'], dt_partition):
dt_move(dt_object, dt_partition, jobs)
else:
if project.verbose: print ('DT Partition Exists:', dt_partition)
for count, job in enumerate(jobs):
print('Waiting For Job: %d of %d' % (count + 1, len(jobs)))
job_wait(project.task['to']['auth'], job)
def archive():
if project.verbose:
print('ARCHIVE')
day = project.date - timedelta(days=abs(project.task['days']))
for object in object_list(project.task['auth'],
project.task['storage']['bucket'] + ':' +
project.task['storage']['path'],
files_only=True,
raw=True):
object_day = datetime.strptime(object['updated'],
'%Y-%m-%dT%H:%M:%S.%fZ').date()
if object_day <= day:
if project.task.get('delete', False) == False:
if project.verbose:
print('ARCHIVING FILE:', object['name'])
object_move(
project.task['auth'],
'%s:%s' % (object['bucket'], object['name']),
'%s:archive/%s' % (object['bucket'], object['name']))
else:
if project.verbose:
print('DELETING FILE:', )
object_delete(project.task['auth'],
'%s:%s' % (object['bucket'], object['name']))
def storage_list(account):
path = '%s:' % account.get_bucket(full_path=False)
project.initialize(_project=CLOUD_PROJECT, _service=CLOUD_SERVICE)
try:
for filename_storage in object_list('service', path, files_only=True):
yield Storage(filename_storage)
except:
pass # if no bucket then skip ( maybe not set up yet )
def archive():
if project.verbose: print('ARCHIVE')
day = project.date - timedelta(days=abs(project.task['days']))
if 'storage' in project.task:
for file_name in object_list(project.task['auth'],
project.task['storage']['bucket'] + ':' +
project.task['storage']['path'],
files_only=True):
file_day = parse_yyyymmdd(file_name)
if file_day and file_day <= day:
if project.task.get('delete', False) == False:
if project.verbose: print('ARCHIVING FILE:', file_name)
object_move(project.task['auth'], file_name,
file_name.replace(':', ':archive/'))
else:
if project.verbose: print('DELETING FILE:', file_name)
object_delete(project.task['auth'], file_name)
def dt():
if project.verbose: print "DT TO TABLE", project.task['to']['table']
delimiter = '\n'
disposition = 'WRITE_TRUNCATE'
# loop all dt files to match pattern
path = '%s:%s' % (project.task['from']['bucket'],
project.task['from']['path'])
for dt_file in object_list(project.task['auth'], path, files_only=True):
# decompression handler for gzip ( must be outside of chunks as it keeps track of stream across multiple calls )
gz_handler = zlib.decompressobj(32 + zlib.MAX_WBITS)
# sliding view of data flowing out of decompression, used to buffer and delimit rows
first_row = True
view = ''
# loop all chunks of file, decompress, and find row delimiter
for data_gz in object_get_chunks(project.task['auth'], dt_file):
view += gz_handler.decompress(data_gz.read())
if first_row:
end = view.find(delimiter)
schema = dt_schema(view[:end].split(','))
view = view[(end + 1):]
first_row = False
end = view.rfind(delimiter)
io_to_table(project.task['auth'], project.id,
project.task['to']['dataset'],
project.task['to']['table'], BytesIO(view[:end]),
'CSV', schema, 0, disposition, False)
disposition = 'WRITE_APPEND'
view = view[min(end + 1, len(view)):]
