About RiC-O Converter

Why RiC-O Converter?

The Archives nationales de France (ANF) have been interested in entity/relationships models and graph technologies since 2013, one of the reasons being that they already have authored a significant, and growing, number of authority records (on corporate bodies, persons and families that created or accumulated the archival fonds held by the institution) that were linked to each other and to the descriptions of the archives themselves. This in essence constitutes a very dense oriented graph, whose relations are not really displayed, and cannot be queried and processed in the ANF current information system. The ANF also wanted to connect these metadata with other metadata sets created by other institutions. Linked Data technologies thus seemed to be a possible solution to meet these needs.

The ANF also have been actively involved in the development of Records In Contexts standard (both the Conceptual Model and its OWL implementation, RiC-O) since the beginning, in 2012, when the ICA EGAD group was formed. As the first version of RiC-CM and beta versions of RiC-O were being written, it was very important to test RiC and its ability to represent existing metadata, and to learn from these tests, both for improving RiC and for thinking of future real world implementations.

The ANF first built a qualitative proof of concept, with two other institutions, the French National Libray (BnF) and a subdivision of the French ministry of Culture (the Service ministériel des Archives de France, SIAF), and a private company (Logilab). The target was to demonstrate that it is possible to convert existing archival metadata sets into RDF datasets conforming to RiC-O (in 2017-2018 it was an early version of the ontology), to interconnect the RDF datasets coming from several institutions and to explore and visualize them using new methods. The proof of concept, named PIAAF (Pilote d’Interopérabilité pour les Autorités archivistiques françaises) was released in February 2018 (https://piaaf.demo.logilab.fr), with a tutorial, and a report on the methodology followed and outcomes. Not only were the results obtained of high quality but the partners were also able to assess the improvement this solution offers in terms of accuracy and search possibilities.

However, PIAAF proof of concept did not take into account the huge quantity of archival metadata created and maintained by the ANF (about 32.000 EAD 2002 finding aids and 16.000 EAC-CPF authority records for now), nor the variety of their structure and content. The datasets that had been selected for the project were small (276 authority records and 38 finding aids), came from three institutions, and had been very carefully checked, so that they conformed to some common rules, which resulted to a quite homogeneous corpus. Besides, the reference ontology, RiC-O, changed a lot since the end of 2017, and its first public release, dated December 2019, is very different from the version used two years earlier.

RiC-O Converter main features

In order to move forward to a far larger scale, the ANF decided in 2018 to develop a tool for converting the whole of their archival metadata, that would have, when finished (thus by the beginning of 2020) the following features:

output RDF datasets conforming to the latest official version of RiC-O;
be efficient (take into account any component of the ANF XML files, in the best possible way) and fast (the conversion process should last not more than a few minutes);
be autonomous, independent from the ANF archival information system, easy to install and process by any archivist on a personal computer;
be configurable;
be precisely documented;
be open source and released under a free license, to that it can be modified by any institution or person which would need to do so.

The ANF therefore worked for a year with a small private company specializing in semantic web, information management and knowledge engineering, Sparna. RiC-O Converter is the result of this project. The Department of digital innovation of the French Ministry of Culture funded and supported this project, as part of the semantic roadmap of the ministry for which it is responsible.

The ANF now have converted all their EAD and EAC-CPF files metadata to RDF; and they are updating and enhancing the RDF resulting files. They also have converted to RDF (the main reference models being RiC-O and SKOS) the whole of their controlled vocabularies and lists of places. The resulting RDF datasets (authority records and vocabularies) are now available in a public repository on GitHub, and are evolving. As they wish to add some semantic modules to their archival information system, they are now exploring some leads through several projects, in order to define a global Linked Open Data strategy.

RiC-O Converter was developed with and for the ANF and does not aim to be a generic tool. It therefore has some limits. However, as it has the features listed above, it may be a good starting point for any project that would need to get good quality RDF/RiC-O datasets from existing EAD or EAC-CPF files.

Of course, French archives, which have encoding practices quite similar to those of the ANF, could be the first institutions interested. However, EAD 2002 and EAC-CPF are now widely used worldwide, by many archive services and by portals or aggregators, as storage and exchange formats. If different practices can be observed there, these institutions share at least the same XML reference models, within which certain elements or attributes are systematically used. We therefore believe that RiC-O Converter can be useful to many institutions other than the ANF. Note that the conversion of EAD files is done separately from that of EAC-CPF files. So if you have ‘only’ EAD files or ‘only’ EAC-CPF files, you can use it. In addition, the software, placed under CeCILL-B license (equivalent to MIT license), can be modified to be adapted to a specific project, and also widely redistributed. It has been precisely documented in English (see the EAD to RiC-O and EAC-CPF to RiC-O mappings, and the unit tests) to facilitate these adaptations. A good XSLT developer, knowing the formats of the source metadata, and having taken a little time to understand the essentials of RiC-O, can easily adapt RiC-O Converter. Finally, the ANF and Sparna are continuing to develop it. After RiC-O Converter 2, they will most probably work on a v3 of the tool, that will conform to RiC-O 1.0 (the first official, stable, version of the ontology, which will be released by the beginning of October 2023). We hope that a larger community will be formed to continue the developments.

More information

An article in English about RiC-O Converter was accepted in January 2023 by the Journal on Computing and Cultural Heritage (JOCCH) and is now published in the Volume 16, Issue 3 (September 2023) of this journal: see https://doi.org/10.1145/3583592. This article quotes the ANF Sparnatural demonstrator (https://sparna-git.github.io/sparnatural-demonstrateur-an/), which provides a friendly interface for exploring and viewing a knowledge graph of about 59 million RDF/RiC-O triples produced using RiC-O Converter (the dataset is there: https://github.com/ArchivesNationalesFR/Sparnatural_prototype_data).

You can access some slides (in French) on RiC-O Converter (presented on January 28, 2020), here.

If you or your institution are a member of ICA, you also can find an article on RiC-O Converter in Flash (the ICA biannual newspaper, dated April 2020) n° 39.

About the ANF projects: the following presentation, made in Lausanne (Switzerland) in December 2022, in French, provides an update on all the projects in progress: slides; video recording. You can also read this article (in French) written by Florence Clavaud: Transformer les métadonnées des Archives nationales en graphe de données : enjeux et premières réalisations, in Les Archives nationales, une refondation pour le XXIe siècle, La Gazette des Archives, n°254 (2019-2), Association des Archivistes Français, Paris, 2019, p. 59-88.