def init(dataset_name):
"""
Initialize a new dataset at the current dir.
Then run the upload command to copy all the files in this
directory to FloydHub.
floyd data upload
"""
dataset_obj = DatasetClient().get_by_name(dataset_name)
if not dataset_obj:
namespace, name = get_namespace_from_name(dataset_name)
create_dataset_base_url = "{}/datasets/create".format(
floyd.floyd_web_host)
create_dataset_url = "{}?name={}&namespace={}".format(
create_dataset_base_url, name, namespace)
floyd_logger.info(
("Dataset name does not match your list of datasets. "
"Create your new dataset in the web dashboard:\n\t%s"),
create_dataset_base_url)
webbrowser.open(create_dataset_url)
name = click.prompt(
'Press ENTER to use dataset name "%s" or enter a different name' %
dataset_name,
default=dataset_name,
show_default=False)
dataset_name = name.strip() or dataset_name
dataset_obj = DatasetClient().get_by_name(dataset_name)
if not dataset_obj:
raise FloydException(
'Dataset "%s" does not exist on floydhub.com. Ensure it exists before continuing.'
% dataset_name)
namespace, name = get_namespace_from_name(dataset_name)
data_config = DataConfig(name=name,
namespace=namespace,
family_id=dataset_obj.id)
DataConfigManager.set_config(data_config)
floyd_logger.info(
"Data source \"{}\" initialized in current directory".format(
dataset_name))
floyd_logger.info("""
You can now upload your data to Floyd by:
floyd data upload
""")
def init(dataset_name):
"""
Initialize a new dataset at the current dir.
After init ensure that your data files are in this directory.
Then you can upload them to Floyd. Example:
floyd data upload
"""
dataset_obj = DatasetClient().get_by_name(dataset_name)
if not dataset_obj:
create_dataset_base_url = "{}/datasets/create".format(
floyd.floyd_web_host)
create_dataset_url = "{}?name={}".format(create_dataset_base_url,
dataset_name)
floyd_logger.error(
("Dataset name does not match your list of datasets. "
"Create your new dataset in the web dashboard:\n\t%s"),
create_dataset_base_url)
webbrowser.open(create_dataset_url)
return
data_config = DataConfig(name=dataset_name, family_id=dataset_obj.id)
DataConfigManager.set_config(data_config)
floyd_logger.info(
"Data source \"{}\" initialized in current directory".format(
dataset_name))
floyd_logger.info("""
You can now upload your data to Floyd by:
floyd data upload
""")
def add(source):
"""
Create a new dataset version from the contents of a job.
This will create a new dataset version with the job output.
Use the full job name: foo/projects/bar/1/code, foo/projects/bar/1/files or foo/projects/bar/1/output
"""
new_data = DatasetClient().add_data(source)
print_data([DataClient().get(new_data['data_id'])])
def add(source):
"""
Create data for current dataset from a given source, for example: foo/projects/bar/1/output
"""
new_data = DatasetClient().add_data(source)
print_data([DataClient().get(new_data['data_id'])])