datalab Federation

This repository contains an opt-in list of active datalab instances. Registered your instance with this federation allows you to make use of certain shared features.

Use cases

Persistent link resolver for your datalab items (e.g., samples)

The service running at https://purl.datalab-org.io can be used to redirect any requests for prefixed IDs to the correct registered datalab instance. For example, a request for purl.datalab-org.io/example:AABBCC would be redirected to the correct URL for the example provider, if registered.

This service is intended to be used when creating QR codes or other labels with datalab, to avoid labels becoming obsolete if the instance URL or underlying API changes (due e.g., to a change in hosting).

These links can also be shared with selective permissions, to allow the holder to see e.g., safety information or compositions, without needing to log in.

Planned features:

• Streamlining the sharing of data between different datalab instances in the future,
• Enable archival of samples beyond the life of a datalab instance. For example, if a group stops using datalab, any important information regarding their samples (composition, safety information) that should remain accessible to someone physically holding the sample need to can be preserved under the purl.datalab-org.io URL.

Future use-cases

• Stats dashboard and monitoring
• OSS plugin registry
• Shared infrastructure:
  • pooled resources for offsite backups and data storage,
  • shared MongoDB replica sets with oplog enabled

Registration process

To register your datalab instance with the federation, please open a pull request to this repository that does the following:

1. Choose a unique ID or <prefix> for your datalab instance. These should be short (though we can be somewhat flexible), containing lowercase letters and hyphens only.
2. Create a single YAML file under the ./src/instances directory called ./src/instances/<prefix>.yaml with the following format.
3. Populate this file with the metadata of your instance, following the schema at ./schema.yaml. You can use an existing instance (e.g., demo.yaml) as a template.
4. If you are familiar with the process, you can optionally validate this locally by installing the prerequisite tools and running make validate, or building the Docker image from the ./Dockerfile.
5. Create your pull request to this repository with the new file; it will be automatically validated by the GitHub Actions CI and then reviewed by an admin.

Caution

Once accepted, your datalab instance URLs will be public, so ensure you have set up your account registration process to be as permissive/restrictive as you desire.