Welcome to the official GitHub page of the Graph-Based Archival Description project. The project is brought to you by a team from the Archives of Ontario and University of Toronto and co-funded from a Canadian federal grant.
The grant was completed in December 2025. One of its deliverables is DrawRDF – a Draw.io plugin that adds an option to export diagrams to Resource Description Framework (RDF) graphs. It is available publicly and for free here: https://gbad-project.github.io/drawio/src/main/webapp/?p=rdf
DrawRDF is included in the ICA/EGAD moderated Records in Contexts Resource List: https://ica-egad.github.io/RiC-ResourceList/resource-details/42.html
A 5-minute video demo of basic functionality is available on YouTube: https://www.youtube.com/watch?v=LaUAY8NCPqY
Additional documentation for DrawRDF is available here: https://drawrdf.readthedocs.io/
A whitepaper is currently being prepared for publication. We will post a notification here as soon as it is available.
Please feel free to contact us if you have any questions.
Thank you for your interest in our project!
Download the slide deck PDF here: https://github.com/gbad-project/.github/raw/refs/heads/main/Graph-Based_Archival_Description_-_ACA_Presentation_-_2025-06-10.pdf
April 2025 Update: To allow archivists, librarians and others to replicate this work within their own institutions, Bachelor of Information students Thomas Fox, Harrison Huang and Russell Luchin created a toolkit for the Graph-Based Archival Description used at the Archives of Ontario.
Download the Toolkit PDF here: https://github.com/gbad-project/.github/raw/refs/heads/main/GBAD_Toolkit_Fox_Huang_Luchin_2025-04-10.pdf
License: CC BY 4.0 International
Note: As of June 2, 2025, the Toolkit was found to require additional Python programming to be reproduced. Please open an issue if you are reading this and need help, or reach out using the contacts indicated on the project’s GitHub page at: https://github.com/gbad-project
The Archives of Ontario, the GLAM Incubator and Profs. Anastasia Kuzminykh and Shion Guha at the University of Toronto have collaborated to create a proof of concept that explores the possibilities and potential benefits of a graph-based data model for archival description, using the International Council on Archives’ new Records in Contexts (RiC) standard and Linked Open Data (RDF data format) (Introductory YouTube video). The traditional archival finding aid, with its static and inflexible hierarchies, fails to represent the complexity and nuanced reality of record creation, accumulation, use and re-use over time.
Recent efforts to overcome this limitation have turned to entity-relationship based models developed by the graph theory of knowledge representation, with its infinitely flexible and extensible subject-predicate-object expressions of data (semantic triples). The overall objective of this project is to develop a data model, to select and model a small-scale sample of data drawn from the Archives of Ontario’s datasets, and to thereby test the possibilities and potential benefits for internal staff and public end-users. This work represents a paradigm shift for archival description: it promises a vastly improved ability to represent the reality of record creation, to manage the complexity of digital metadata, to allow machine-readable encoding of meaning that enables complex logical inferences and search capabilities, and to connect currently siloed datasets with related data across the world. In short, this proposed project will break new ground in the Canadian archival sector.
Feature Article. https://ischool.utoronto.ca/news/a-glam-makeover-for-the-archives-of-ontario/
YouTube Video. https://youtu.be/3ZtTppHhyN0
This page is in development. Here are the main repositories you can access through this GitHub page:
Our fork of Richard Williamson’s Draw.io parser with added support for modelling RML graphs. https://github.com/gbad-project/records_in_contexts_draw_io_parser
Our fork of Draw.io itself, with a custom plugin for RDF export. https://github.com/gbad-project/drawio/blob/gbad/src/main/webapp/plugins/rdfexport/README.md
Sparnatural fork and demo. https://github.com/gbad-project/gbad-project.github.io
Our fork of a Visual Studio Code Turtle RDF code autocomplete extension. https://github.com/gbad-project/turtle-vocab-autocomplete
For a simple and lightweight URI dereferencer, supported by a Python/Oxigraph-backed YASGUI instance and suitable for ultralightweight VPS deployments (under 2 GB RAM!), reach out using our contact email.
Les données RDF interrogeables par la présente application web ont été produites par
les Archives nationales. Il s’agit d’informations publiques ; l’usager dispose d’un
droit non exclusif et gratuit de libre réutilisation de ces données à des fins
commerciales ou non, dans le monde entier et pour une durée illimitée. Il doit
accompagner chaque rediffusion des informations de l’indication précise de l’origine
des métadonnées : « Archives nationales (France), date de ces métadonnées (août 2022),
métadonnées du démonstrateur Sparnatural ».
Voir à ce sujet la page : Réutilisation des données publiques sur le site des Archives nationales.
Les textes constituant la documentation et les autres documents de mise en contexte publiés par l’équipe projet au sein du site web sont mis à disposition selon les termes de la licence Creative Commons « Attribution – Pas d’Utilisation Commerciale – Partage dans les Mêmes Conditions ; 4.0 International (CC-BY-NC-SA 4.0) » (http://creativecommons.org/licenses/by-nc-sa/4.0/deed.fr).
Vous pouvez télécharger les données RDF depuis le dépôt public dans lequel nous le gérons sur GitHub : https://github.com/ArchivesNationalesFR/Sparnatural_prototype_data.
Le démonstrateur a été réalisé par la société Sparna et par le Lab des Archives nationales. Le démonstrateur utilise une version de l’éditeur open source de requêtes Sparnatural développée dans le cadre d’un projet associant les Archives nationales, la BnF et le département du Numérique pour la transformation des politiques culturelles et l’administration des données du ministère de la Culture.
Le triplestore RDF dans lequel sont stockées les données est une instance du logiciel libre GraphDB Free, que la TGIR Huma-Num du CNRS héberge sur sa grille de services.
Nous remercions en outre les participants aux ateliers des 16 et 22 novembre 2021 pour leur contribution active : Nathalie Abadie, Clément Arnaud, Laurence Croq, Jean-Michel Duchemin, Chloé Fize, Jasmine Gherram, Julien Le Magueresse, Marie-Véronique Leroi, Xavier Levoin, Jean-François Moufflet, Gaetano Piraino, Michel Schonn, Antoine Silvestre de Sacy, Vincent Verdese.
Vous souhaitez nous signaler une erreur, nous poser une question ou en savoir plus
sur le projet ?
Contactez-nous en écrivant à le-lab.archives-nationales@culture.gouv.fr !