modelarts-umn/api-ref/source/data_management/data_management_apis.rst

original_name

modelarts_03_0299.html

Data Management APIs

Table 1 Data management APIs

Type	API	Description
Dataset Management	Querying the Dataset List <listdatasets>	Query created datasets by page based on specified conditions.
	Creating a Dataset <createdataset>	Create a new dataset .
	Querying Details About a Dataset <descdataset>	Query details about a dataset, including the dataset name, type, and version name based on the dataset ID.
	Modifying a Dataset <updatedataset>	Modify basic information about a dataset, such as the dataset name, description, version, or labels. The modification takes effect in the entire dataset, including the samples in the dataset.
	Deleting a Dataset <deletedataset>	Delete a dataset based on the dataset ID to release resources.
	Querying Dataset Statistics <liststats>	Query dataset statistics, such as sample statistics, label statistics, or hard examples based on specified conditions.
	Querying the Monitoring Data of a Dataset <getdatasetmetrics>	Query the monitoring data of a dataset within a specified period, such as the number of labeled samples, number of unlabeled samples, and total number of samples at each time point within the period.
Data Synchronization	Synchronizing a Dataset <syncdatasource>	Synchronize data and labels from the dataset input path to the dataset.
	Querying the Status of a Dataset Synchronization Task <syncdatasourcestate>	Query the status of a data source synchronization task based on the dataset ID.
Labeling Data	Updating Sample Labels in Batches <updatesamples>	Label multiple samples in a dataset in batches. Label unlabeled samples. You can use an existing label or create a new label. Add, modify, or delete labels for labeled samples. You can use an existing label or newly added label to modify an original label. Additionally, you can add an existing label or a new label to a sample. This API uses a new label list to overwrite the original one to update the sample label. For example, if an empty label list is used to overwrite the original one, all sample labels will be deleted.
	Querying the Sample List <listsamples>	Query dataset samples by page based on specified conditions.
	Adding Samples in Batches <uploadsamplesjson>	Add samples to a dataset in batches for data labeling.
	Deleting Samples in Batches <deletesamples>	Delete unused samples from a dataset in batches.
	Querying Details About a Sample <describesample>	Query a single sample based on the sample ID, including the sample status and labels.
	Querying Sample Search Criteria <listsearch>	Obtain sample search criteria, such as the label list and attribute key-value pairs of the dataset based on the dataset ID.
Label Management	Querying the Dataset Label List <listlabels>	Query the labels in a specified dataset version.
	Creating a Dataset Label <createlabels>	During dataset labeling, new labels can be created. This function is available only in datasets of the text classification and named entity types.
	Modifying Labels in Batches <updatelabels>	Modify dataset labels in batches. The modification takes effect in the entire dataset, including the samples in the dataset.
	Deleting Labels in Batches <deletelabels>	Delete dataset labels in batches and ask you to determine whether to delete the samples with the labels.
	Updating a Label by Label Name <updatelabel>	Modify a label in a dataset based on the label name. The modification takes effect in the entire dataset, including the samples in the dataset.
	Deleting a Label and the Files with This Label Only <deletelabelandsamples>	Delete a label in a dataset based on the label name and ask you to determine whether to delete the samples with the label.
Importing Data	Querying the Dataset Import Task List <listimporttasks>	Query historical tasks imported to a dataset by page based on the dataset ID.
	Creating a Dataset Import Task <importtask>	Create a dataset import task to import labels and data (such as manifest files and OBS data) from a storage system to the dataset.
	Querying Details About a Dataset Import Task <descimporttask>	Query details about a dataset import task based on the dataset ID and task ID to learn about the data source, import mode, and task status.
Exporting Data	Querying the Dataset Export Task List <getexporttasksstatusofdataset>	Query historical tasks exported from a dataset by page based on the dataset ID.
	Creating a Dataset Export Task <exporttask>	Export certain data as a new dataset or to OBS.
	Querying the Status of a Dataset Export Task <getexporttaskstatusofdataset>	Query details about a dataset export task based on the dataset ID and task ID to learn about the export type, task status, and number of samples.
Publishing a Dataset	Creating a Dataset Labeling Version <createdatasetversion>	Publish the labeled dataset as a new version for model building.
Managing Dataset Versions	Querying the Dataset Version List <listdatasetversions>	Query the versions of a dataset based on the dataset ID to learn about the dataset version evolution.
	Creating a Dataset Labeling Version <createdatasetversion>	Publish a modified dataset as a new version. The modification includes labeling samples, adding samples, and deleting samples in the dataset.
	Querying Details About a Dataset Labeling Version <describedatasetversion>	Query details about a specified dataset labeling version, including the name, description, number of files, and storage path based on the dataset ID and version ID.
	Deleting a Dataset Labeling Version <deletedatasetversion>	Delete a dataset version based on the dataset ID and version ID.