Data Deletion
Autosuggest provides you with flexible options for data deletion, allowing you to tailor the process to your needs. You have the option to manage data by dataset names or by tagging multiple datasets with a customer ID, for easier data deletion. Delete datasets either by name, or by tag.
Deletion is processed asynchronously: the API call returns as soon as the deletion has been accepted, and the underlying examples are removed shortly afterwards.
Delete by Dataset Name¶
To delete specific datasets, you can use the dataset name as a reference. This method is straightforward and effective for removing individual datasets from your system.
DELETE v2/datasets/<dataset-name>¶
https://api.stag.asgt.visma.ai/v2/datasets/dataset-name<dataset-name> should be taken from the body of the create dataset request.
Authorization - Bearer Token¶
Token: [API token]
Body - empty¶
The response contains the deleted dataset.
Delete by Tag¶
Alternatively, you have the option to tag datasets with identifiers, such as a customer ID, for streamlined data deletion. This method proves useful when managing multiple datasets related to a specific customer, as an example. You can define tags at creation of the dataset, by tagging it with e.g. a customer ID.
Deleting a tag removes all datasets carrying that tag, regardless of their type. Dataset names are not considered - only the tag. This streamlined approach is particularly beneficial when, for instance, you need to delete all data related to a particular customer.
DELETE v2/tags/<tag-name>¶
¶https://api.stag.asgt.visma.ai/v2/tags/customer-1Authorization - Bearer Token¶
Token: [API token]
Body - empty¶
Warning
A tag can only be deleted if it matches ^[A-Za-z0-9.][A-Za-z0-9_.>-]*$ - letters, digits, ., _, > and -, starting with a letter, digit or .. Tags may be created with spaces in them, but such a tag cannot be used to delete datasets. Use hyphenated tags such as customer-1 if you intend to delete by tag later.
Remove examples but keep models¶
If you want to clear the examples a dataset trains on without removing the dataset or the models already trained from it, truncate the dataset instead of deleting it.
DELETE v2/datasets/<dataset-name>/examples¶
¶https://api.stag.asgt.visma.ai/v2/datasets/dataset-name/examplesAuthorization - Bearer Token¶
Token: [API token]
Body - empty¶
Existing models remain available for predictions, and future trainings will only use examples added after the truncation.
Retention¶
Datasets also expire data automatically. See Data Retention for the retention window and its default.