The goal of the Uniform Data Access Layer is to:

  • distinguish between the retrieval of data and its processing
  • make data access compatible between providers
  • allow re-use of software packages

There is no single UDAL implementation, only a specification they must follow. This allows for different people to create implementations for the data they provide or for the data they need to access if no one else provided a package already. Moreover, queries are not part of the specification, and implementors are free to choose the queries and optional parameters they wish to support.

UDAL overview

The specification for UDAL implementations describes how the code for an implementation should be. In Python, this is done using abstract base classes and it is available at https://github.com/fair-ease/py-udal-interface .

UDAL implementations can include support for multiple queries and distinct data sources. A connection string provided by the user will set which data sources and queries are available. Access to the data and behaviour of the implementation are defined by configuration items provided by the user.

UDAL implementation and usage example

Named queries are specified using RDF and an initial set of queries is available at https://github.com/fair-ease/dataset-demand-register .