FAIR Data Services Version 1.20.0

FAIR API users can use the built-in /api-docs feature to review the API documentation and (where authorised) execute some API calls. Due to some out of date documentation the API call for GET /dictionaries/{code} was not working - this has now been rectified.

This bug fix addresses an issue with downloading data and resources (where authorised). Due to a misconfiguration of the user interface, an error was preventing downloads.

This bug fix address an API issue whereby a user requesting a list of datasets (metadata) would get an internal server error. In the web user interface this error was not appearing. Investigation of the bug showed that it was caused by the way underlying data sources were being managed with an error being raised if a dataset was orphaned of the original data source. This has now been rectified and metadata returned without error.

FAIR data services can host data as well as metadata. As part of developing the data management features of the product, we are introducing type handling for data. Previously uploaded data was held in PostgreSQL using generic text field types. With this enhancement, we take the data type provided in the dataset metadata and use that as the target field type in the database. As data is uploaded we attempt to convert the incoming (CSV) data into the target field type. If there are problems with the conversion, these are logged as errors.

For security reasons, login sessions are time limited. Some users reported an issue due to session timeouts that was happening when saving dataset metadata and a related issue with resource and data upload. With this change we hope to reduce the times when a user may appear to be logged in but their session has expired and they cannot proceed with saving or uploads. In those cases, we will attempt to prompt the user to log in and update the login session so that they can complete the action.