FAIR Data Services Version 1.26.0

FAIR’s catalogue can be configured in ‘data commons’ mode, where non-registered and non-logged in users can search and view metadata, with limited functionality. A security review highlighted the risk of denial of service attacks on a publicly-accessible website like this. In order to mitigate the risk, we will limit the rate at which anonymous guest users can make requests to the service. Customers remain in control of enabling data commons mode and can additionally request configuration of the rate limiting parameters.

We aim to make it easier to integrate FAIR with other systems in a customer’s ecosystem. We are introducing custom actions for FAIR datasets - providing additional links on the action toolbar in the dataset view. For example, a customer may have another tool that provides visualisation, or analysis or discussion related to a dataset or part of a dataset. Users can then jump directly from the dataset view to that tool.

Administrators can now add up to five custom actions which are available to data owners. They can specify the label, a tooltip and the URL template for the action. In addition, via service desk support, they can request a custom icon to be displayed. Custom actions open a new tab by default.

URL templates can use fields from the dataset catalogue. For example, a URL might require the dataset code (as a unique identifier for the dataset), but other keys could be used such as a DOI (the identifier field). In some cases, a field may require hierarchical key e.g. the name of the publisher is publisher.name. Some examples of URL templates:

https://www.example.com/path/to?dataset={code}
https://www.example.com/path/to/{code}
https://{code}.datasets.example.com
https://www.example.com/path/to/doi-aggregator/?doi={identifier}
https://www.example.com/path/to/?publisher={publisher.name}

Earlier this year, we introduced custom catalogue templates which enable a customer to extend the fields describing a dataset. As part of a phased introduction, we did not enforce compliance with the custom catalogue template at the time of saving catalogues. Going forward, when users create a new dataset catalogue entry or update an existing catalogue entry, its structure and contents will be checked against the default catalogue template set for that instance of FAIR. The user will not be able to save the catalogue if it does not match the required fields.

Note: Currently, only users of the API are able to submit entries with the full custom catalogue. Editing of full custom catalogue will be added to the user interface later in the year.

Based on user feedback, we have improved the API support for dataset listing. The user interface will be updated in a later release. Previously, selecting the icon to ‘Show my datasets’ was too narrowly defined - it will now filter datasets based on sharing permissions. This should allow a user to see all the datasets they have elevated privileges to (e.g. to edit, or to view a private dataset).

FAIR’s data delivery pipeline can now extract data from multiple sources including PostgreSQL, SQL Server and Azure Synapse. These sources are configured by the customer as external data sources per-dataset.

As part of implementing this feature we have also made some improvements to the data delivery pipeline with better handling of table names (dictionary codes) with upper and lower case characters as well as improving the error reports on failed transfers.

Customers building their own platforms on the Aridhia DRE, including FAIR, want to understand usage patterns and how users navigate the DRE. In order to support this requirement, but mindful of user privacy, we have integrated a privacy-preserving web analytics component.

This feature is currently only enabled on implementations where the customer has requested it, is limited in scope to the DRE and customer’s own adjacent services, and requires users to consent to their browsing data being collected for this purpose.

This is a small user interface tweak to standardise icons and icon sizes for service banners.

We fixed a user interface bug with adding URL resources - this minor update should make it easier for users to choose whether to add a file resource or add a URL resource.

We fixed a user interface bug where overlapping components prevented completion of a data access request form when validation errors are found.

Data access requests can be customised using a business process model runtime engine. FAIR provides details for the request to the process which may interact with third party services from data providers, including federated data sources. With this change we have enriched the information provided to make it easier to align references for datasets which may appear in both FAIR and the federated data provider and more details of the dataset catalogue entry.