create_dataset
(**kwargs)¶Creates a new dataset in an Amazon Lookout for Vision project. CreateDataset
can create a training or a test dataset from a valid dataset source ( DatasetSource
).
If you want a single dataset project, specify train
for the value of DatasetType
.
To have a project with separate training and test datasets, call CreateDataset
twice. On the first call, specify train
for the value of DatasetType
. On the second call, specify test
for the value of DatasetType
.
This operation requires permissions to perform the lookoutvision:CreateDataset
operation.
See also: AWS API Documentation
Request Syntax
response = client.create_dataset(
ProjectName='string',
DatasetType='string',
DatasetSource={
'GroundTruthManifest': {
'S3Object': {
'Bucket': 'string',
'Key': 'string',
'VersionId': 'string'
}
}
},
ClientToken='string'
)
[REQUIRED]
The name of the project in which you want to create a dataset.
[REQUIRED]
The type of the dataset. Specify train
for a training dataset. Specify test
for a test dataset.
The location of the manifest file that Amazon Lookout for Vision uses to create the dataset.
If you don't specify DatasetSource
, an empty dataset is created and the operation synchronously returns. Later, you can add JSON Lines by calling UpdateDatasetEntries.
If you specify a value for DataSource
, the manifest at the S3 location is validated and used to create the dataset. The call to CreateDataset
is asynchronous and might take a while to complete. To find out the current status, Check the value of Status
returned in a call to DescribeDataset.
Location information for the manifest file.
The S3 bucket location for the manifest file.
The Amazon S3 bucket that contains the manifest.
The name and location of the manifest file withiin the bucket.
The version ID of the bucket.
ClientToken is an idempotency token that ensures a call to CreateDataset
completes only once. You choose the value to pass. For example, An issue might prevent you from getting a response from CreateDataset
. In this case, safely retry your call to CreateDataset
by using the same ClientToken
parameter value.
If you don't supply a value for ClientToken
, the AWS SDK you are using inserts a value for you. This prevents retries after a network error from making multiple dataset creation requests. You'll need to provide your own value for other use cases.
An error occurs if the other input parameters are not the same as in the first request. Using a different value for ClientToken
is considered a new call to CreateDataset
. An idempotency token is active for 8 hours.
This field is autopopulated if not provided.
dict
Response Syntax
{
'DatasetMetadata': {
'DatasetType': 'string',
'CreationTimestamp': datetime(2015, 1, 1),
'Status': 'CREATE_IN_PROGRESS'|'CREATE_COMPLETE'|'CREATE_FAILED'|'UPDATE_IN_PROGRESS'|'UPDATE_COMPLETE'|'UPDATE_FAILED_ROLLBACK_IN_PROGRESS'|'UPDATE_FAILED_ROLLBACK_COMPLETE'|'DELETE_IN_PROGRESS'|'DELETE_COMPLETE'|'DELETE_FAILED',
'StatusMessage': 'string'
}
}
Response Structure
(dict) --
DatasetMetadata (dict) --
Information about the dataset.
DatasetType (string) --
The type of the dataset.
CreationTimestamp (datetime) --
The Unix timestamp for the date and time that the dataset was created.
Status (string) --
The status for the dataset.
StatusMessage (string) --
The status message for the dataset.
Exceptions
LookoutforVision.Client.exceptions.AccessDeniedException
LookoutforVision.Client.exceptions.InternalServerException
LookoutforVision.Client.exceptions.ValidationException
LookoutforVision.Client.exceptions.ConflictException
LookoutforVision.Client.exceptions.ResourceNotFoundException
LookoutforVision.Client.exceptions.ThrottlingException
LookoutforVision.Client.exceptions.ServiceQuotaExceededException