Get the list of alerts.

Alerts are the notifications to the user about what has gone wrong and how to fix it.

Examples:

• /alerts?end=14&limit=3 End in alert ID 14 and show up to 3 alerts. So it could return alerts 12, 13, and 14.
• /alerts?start=2&limit=3 Start from alert ID 2 and show up to 3 alerts. So it could return alerts 2, 3, and 4
• /alerts?limit=3&reverse=false Get up to 3 alerts and don not reverse the order. By default we reverse the order so it is descending from the latest to the earliest alerts.

Get one alert based on its ID

Resolve an alert.

Get the stats for the frontend left menu to be displayed to the user.

Get the list of datasets.

Creates new datasets. The ids and creation dates will be ignored in the request. table_name and schema_name and name are required. For the full list of what is required and what is ignored, take a look at the DataSetPostRequestSchema

Example request:

{ "name": "new", "table_name": "table1", "schema_name": "schema1", "settings_model": { "boolean_true": [ "blah" ] } }

Modify a dataset using a PATCH query.

Example

PATCH /datasets

body

{
    'id': 1,
    'version_id': 1,
    'settings_model': {
        'id': 1,
        'version_id': 2,
        'dollar_to_cent': true,
    },
    'name': 'foo',
}

In this example we are modifying the dataset ID 1 to have a new name: foo. We are slow modifying its associated settings_model to have the dollar_to_cent value of True.

Get a dataset object with default values filled up so it can be used as a basis by the user to create a new dataset by doing a POST query.

Get a detail view of a dataset including the associated settings model.

Note: There is a 1:1 relationship between datasets and settings_models. The reason they are separate is that a settings_model's change can potentially cause a training job. Because we want to reflect any changes to the settings in the output of the training job which can ultimately cause changes in the database schema.

However the configs in the dataset itself are not anything that can cause a retraining.

Get the history of the dataset

Get the row stats for a dataset.

Currently we only return the bad_rows_count which are the count of the rows of the dataset that are in the quarantine table.

Make a copy of the configurations of a dataset

Creates new datasets. The ids and creation dates will be ignored in the request. table_name and schema_name and name are required. For the full list of what is required and what is ignored, take a look at the DataSetPostRequestSchema

Example request:

{ "name": "new", "table_name": "table1", "schema_name": "schema1", "settings_model": { "boolean_true": [ "blah" ] } }