Regulatory Knowledge Graph - Réseau de connaissances reglementaire
This project is a demonstration project, serving as a model for methods to collect and link data from a variety of sources into a single RDF model, then write that model as a packed Turtle (TTL) file.
To demonstrate that the output format need not be restricted to RDF native formats, the project emits a SQLite 3 packed and optimized data file as well. The file meets the conditions to be used client-side with phiresky's WASM-based SQLite3 implementation, as well as any of the other native implementations out there.
Because the business domain chosen (Canadian Federal Regulations) is text-rich, the same process take the opportunity to create a text index of the instruments it assembles in a simple JSON format that just happens to be the same format expected by Apache SOLR. A simple example of a SOLR schema is inclded in the root of the repository.
As of March 4, 2022, this project has completed its development goals. Maintenance is limited to security updates until such time as the steering committee approves additional development. Interested parties are encouraged to communicate with [email protected] for briefings, expressions of interest, etc. As permissively-licensed OSS, community-contributed PRs are always welcome!
Building this project requires Java 17 and Maven to be on your command line path.
mvn clean install
Building can be completed with Docker 18.09 as it uses BuildKit
DOCKER_BUILDKIT=1 docker build --file Dockerfile --output target .
Following the build, the output models can be found in target/out.ttl, target/out.sqlite3, and target/out.json.
To build the project with Docker, use version 18.09 or higher as it uses BuildKit
cd into the front-end folder then run DOCKER_BUILDKIT=1 docker build --file Dockerfile --output build . to build the project.
Following the build, the static HTML files will be output in the front-end/build folder.
For more information on running the front end, see the Read Me.
Information about the RDF schema used can be found in the schema reference.
The tooling is rigged in such a way that extending the model in your local environment is as easy as possible.