“Director’s cut” is a new bimonthly blog series launched on the OpenCitations blog to share the Director’s point of view on the hidden work happening within our infrastructure.
In our roadmap, informed by a community survey we conducted to address the needs of our users (and potential users), we promised to focus, among other things, on strengthening backend development to foster better semantic interoperability with other research infrastructures. Indeed, such a topic is gaining momentum in recent months and has generated a lot of useful discussion within the community interested in open scholarly infrastructures (OSIs) – e.g. with the last and recent event about ORION-DBs. As anticipated in our annual report, we have spent significant effort over the past two years reengineering OpenCitations entire technical infrastructure to better comply with the Principles of Open Scholarly Infrastructure (POSI).
Among the results of our efforts, we have worked on devising a general framework, introduced in a paper, to define the architecture of the computational unit of an OSI grounded in the POSI, defining what we think are the core principles and technical features that should be guaranteed – spoiler: separation of operational domains, orchestration, scalability, observability, automation, and workflow portability. While useful for informing and documenting the choices we made in this context, sharing this framework also helps other OSIs in strengthening their scholarly sovereignty. To enable OSIs and other actors to continue the discussion on the content of the paper introducing the framework, we have created a space on PREreview to favour an open discussion.
The new OpenCitations technical infrastructure has allowed us to improve semantic interoperability with other research infrastructures by adopting the Scientific Knowledge Graph – Interoperability Framework (SKG-IF). SKG-IF is an extensible framework, created in the context of an RDA working group, that enables the exchange of data about core research entities included in scientific knowledge graphs, thus providing a standardised approach to model a minimal set of scholarly data and enabling one to mash up data from different sources to support diverse applications. In particular, SKG-IF includes a data model, expresses data in JSON-LD, and provides an OpenAPI specification for enabling requesting, searching, and filtering scholarly data.
Over the past week, we released the OpenCitations implementation of the SKG-IF API, now available in production. This new API covers requests for all the entities we describe in our data (research products, persons, organisations, and venues). In contrast, other SKG-IF entities (grants, data sources, and topics) may be available in other OSIs that enable access to their data via new SKG-IF API endpoints. From a technical perspective, the OpenCitations SKG-IF API uses the new version of RAMOSE, a tool developed and maintained by OpenCitations to create REST APIs over SPARQL endpoints, which also provides a dedicated SKG-IF module (briefly introduced in a new paper and in its documentation) to simplify adoption by RDF scholarly data sources that want to provide an SKG-IF API for accessing their data.
The adoption of such technologies enables, in principle, the creation of a federation of distributed open research information sources that, by adopting SKG-IF, can provide their data (which can be expressed, internally to an infrastructure, as knowledge graphs, stored in relational databases, etc.) without the need to change their internal data model and technologies. What is needed is an SKG-IF Open REST API endpoint that maps the internal data model into SKG-IF. This choice preserves each OSI’s local autonomy and sovereignty, which is usually maintained by independent organisations with their own governance, and avoids costly redesign of established systems.
At OpenCitations, we are actively involved in the GRAPHIA project, where we are participating in an experiment that aims to create a federated infrastructure for SSH knowledge graphs, using SKG-IF as the interoperability contract for the federation. We will share more about the results of this experiment in one of the upcoming issues of this blog series.