Tools & Software

The ReactAI research agenda is supported by an evolving stack of open-source software, semantic infrastructures and experimental tools. These components operate at different layers of the knowledge-production process: some provide persistent knowledge infrastructure, others support interoperable data representation and transformation, while the newest tools address evidence processing, semantic projection and human review.

Knowledge infrastructure

Wikibase

Wikibase provides a collaborative knowledge-base environment for creating and maintaining structured, linked data. We use it particularly where cultural institutions and communities need an accessible environment for curating entities, identifiers, relationships and contextual knowledge without requiring direct interaction with RDF infrastructure.

Sampo-UI

Sampo-UI is a framework for developing user interfaces for semantic portals (Ikkala et al., 2022). The underlying ideas of the framework are based on the “Sampo” model that contains a sustainable business model for publishing and sharing Linked Data based on a shared ontology infrastructure, the idea of providing the end user with multiple application perspectives to the Linked Data, and a two-step usage cycle based on faceted search combined with ready-to-use data-analytic tooling (Hyvönen, 2023). The framework draws from longstanding research of the Semantic Computing Research Group (SeCo) since 2002 on developing a national Semantic Web infrastructure in Finland and semantic portals, which has resulted in publishing a series of “Sampo portals”, mostly related to the Cultural Heritage domain, that have had millions of end users on the Web.

The Sampo-UI framework makes it possible to create highly customizable and responsive user interfaces that satisfy the requirements for such portals using current state-of-the-art JavaScript libraries and data from SPARQL endpoints, while saving substantial coding effort. Sampo-UI is published on GitHub under the open MIT License and has been utilized in several internal and external projects.

Open data infrastructure

MariaDB

MariaDB provides conventional relational database infrastructure where structured operational data are most appropriately maintained in relational form.

Apache Jena Fuseki

Apache Jena Fuseki provides RDF storage and SPARQL services for semantic representations. It supports the publication, querying and exchange of graph-based knowledge while remaining independent from the interfaces through which that knowledge is created or reviewed.

Data & intermediate representations

Different computational operations require different representations of the same underlying evidence and knowledge. Our software therefore deliberately uses relational tables, RDF graphs and other intermediate representations according to the requirements of the operation rather than treating any single technical format as the universal representation.

dataset

The aim of the dataset package is to make tidy datasets easier to release, exchange and reuse. It organizes and formats data frame R objects into well-referenced, well-described, interoperable datasets into release and reuse ready form.

Increase FAIR use of your datasets: Offer a way to better utilise the utils:bibentry bibliographic entry objects and working with the ROpenSci package RefManageR extending their fields of the Dublin Core and DataCite standards, and making them detachable from the data, i.e., including the bibliographic entries into the attributes of a data frame-like object. See for more information the Bibentry for FAIR datasets vignette.

2.Interoperability outside R: Extending the haven_labelled class of the tidyverse for consistently labelled categorical variables with linked (standard) definitions and units of measures in our defined class; this enables to share exact definitions, units of measures across computers and systems, and increasing the interoperability of the data set from an R data.frame to any standardised statistical or library system. 3. Tidy data tidier, richer: Offering a new data frame format, dataset_df that extends tibbles with semantically rich metadata, ready to be shared on open data exchange platforms and in data repositories. This s3 class is aimed at developers and we are working on several packages that provide interoperability with SDMX statistical data exchange platforms, Wikidata, or the EU Open Data portal. Read more in the Create Datasets that are Easy to Share Exchange and Extend vignette. 4. R+RDF=global interoperability: The From R to RDF vignette shows how to leverage the capabilities of the dataset package with rdflib, an R-user-friendly wrapper on ROpenSci to work with the redland Python library for performing common tasks on rdf data, such as parsing and converting between formats including rdfxml, turtle, nquads, ntriples, and trig, creating rdf graphs, and performing SPARQL queries.

install.packages("dataset")
remotes::install_github("dataobservatory-eu/dataset", build = FALSE)

The current version of the dataset package is in an early, experimental stage. You can follow the discussion of this package on rOpenSci.

Read more or browse the source code via the dataset website.

fscontext

fscontext treats files and filesystem structures as sources of contextual evidence. It supports the acquisition and organisation of observations about digital objects before those observations are incorporated into richer semantic knowledge.

fscontext provides a provenance-aware contextual reconstruction framework for file systems and related digital resource collections.

The package creates reproducible observational snapshots of files, repository structures, and related operational resources, and supports their contextual abstraction, semantic stabilization, and reconstruction-oriented analysis.

Installation

# CRAN release
install.packages("fscontext")

# Latest development version
pak::pak("dataobservatory-eu/fscontext")

The package includes five introductory vignettes that follow the typical fscontext workflow from filesystem observation to semantic stabilisation.

Introduction to fscontext introduces filesystem observations, reproducible snapshots, and contextual reconstruction.
From Filesystem Observation to Human Review demonstrates how Betwixt provides intermediate relational representations for reproducibly reviewing and revising candidate semantic and provenance assertions generated from fscontext observations.
Observing File Organisation demonstrates how repeated filesystem observations can reconstruct ordinary human curation activities such as organising imported photographs into contextual collections.
Working with Record Sets demonstrates how observational data can be transformed into provenance-aware recordset_df objects inspired by the Records in Contexts (RiC) conceptual model.
Prelabelled values and semantic stabilisation introduces lightweight semantic enrichment, rulebooks, and human-in-the-loop refinement workflows.
Organising Evidence with Structural Aggregations demonstrates how structural aggregation metadata can identify potentially informative objects and candidate Record Sets across folders, ZIP archives, and WACZ packages.

Together these vignettes introduce the observational, contextual, semantic, and analytical layers of the package.

betwixt

Betwixt projects semantic assertions and their evidence into relational forms suitable for computation and human review. It provides long, wide, dual-long and dual-wide representations while preserving the identities and relationships required to return reviewed results to semantic knowledge systems.

Human-centred review

review

review supports explicit human review of computationally generated candidate knowledge. It separates candidate generation from professional judgement and records review outcomes so that consequential semantic or institutional commitments do not arise silently from computational output.