Skip to main content


A dataset is an entity that allows you to logically aggregate data in order to later be analyzed or used for machine learning workflows.

Example dataset

On the dashboard, a dataset is shown like the one above. As you can see, every dataset starts with the prefix "ds" followed by a numeric id. To access to a dataset, you must use the entire id (prefix included).

Create a dataset#

First of all, to create a dataset, you need to click the button shown below.

In order to create a dataset, go to the dashboard and click the button that says "Create Dataset".

Once the modal appears, a name and description must be provided in order to distinguish it from others. Once you fill the name and description, press the button Create dataset.

Create dataset modal

After that, the dataset is ready to be used.

Append to dataset#

In the sethealth API exists a module called dataset. This module provides a method get which retrieves a Dataset instnce to handle the dataset created. Once you have the instance, you can easily upload assets with the append method.

import * as sethealth from "@sethealth/core";
const myDataset = await sethealth.dataset.get("ds-5681747056394240");
const opts: DatasetOptions = {
caseID: "<case-id>";
entries: [{
type: "<type>";
payload: "<payload>";
size: `<size>`;

The append method, as you can see above, needs some options. These options are:

  • caseID: (optional) Identier of the case. If not filled, a new identier would be generated.
  • entries: The data that will be appended to the dataset. Every entry has the following fields:
    • type: Type of the entry. For example "tomography" or "csv".
    • payload: The source of the data appended. For example an URL to download the entry data.
    • size: (optional) The approximate size of the file that will be downloaded.