Download Datasets

STRIDE provides a CLI command to download datasets from remote repositories. This guide shows how to download pre-configured datasets as well as custom datasets from GitHub.

Prerequisites

• STRIDE installed and available in your environment

• For private repositories: GitHub CLI (gh) installed and authenticated

List Available Datasets

To see the known datasets available for download along with their available versions:

$ stride datasets list-remote

This will display each dataset’s name, repository, subdirectory, description, and available versions. Datasets may also have an associated test dataset (shown as test_subdirectory) which is automatically downloaded alongside the main dataset.

Download a Known Dataset

To download a known dataset to the default location (~/.stride/data or STRIDE_DATA_DIR):

$ stride datasets download global

This single command downloads both the full global dataset and the smaller global-test subset from the same release archive. The test dataset enables faster iteration during development.

Specify a Data Directory

Use the -d or --data-dir option to download to a specific location:

$ stride datasets download global -d ./my_data

Alternatively, set the STRIDE_DATA_DIR environment variable for a persistent default:

$ export STRIDE_DATA_DIR=/path/to/data
$ stride datasets download global

Specify a Version

By default, the latest release is downloaded. To download a specific version:

$ stride datasets download global -v v0.2.0

Download from a Custom Repository

To download a dataset from any GitHub repository, use the --url and --subdirectory options:

$ stride datasets download --url https://github.com/owner/repo --subdirectory data

Note

The --subdirectory option is required when using --url.

Private Repository Authentication

For private repositories, STRIDE automatically uses your GitHub CLI authentication. Ensure you are logged in:

$ gh auth status

If not authenticated, run:

$ gh auth login