Knowledge graph powered information discovery based on keyword associations.
disco-graph can be deployed with a single command on Linux and MacOS if docker and docker-compose are installed:
Using the shell navigate to the deploy/docker directory and run the following command:
bash disco-graph.bash startTo stop it run:
bash disco-graph.bash stopA more in depth description of the deployment can be found in the Deployment Guide.
The Development Guide details how to set up your machine to contribute to disco-graph.
The disco graph application consists of three major components: The Knowledge Graph, the Backend services and the Frontend. A short description of these components is given in the following paragraphs. For a more in-depth description take a look at the guides to each if the individual components:
Knowledge Graph
graph-connector
UI
The Knowledge Graph is the heart of Disco Graph, as it serves as main data-structure containing the relationships between Publications, Keywords and further information. This datastructure allows for efficient retrieval of relevant information. The Knowledge Graph is accessed through the Backend. The Knowledge Graph consists of RDF triples that are stored in an Apache Jena Fuseki SPARQL Server.
The Frontend allows users to discover the knowledge hidden in the knowledge graph. It combines features to search, view and edit information about publications and provides an interface for managing users and roles through the admin portal.
The Frontend is a webapp based on the Angular Framework, using the Angular Material component library.
The backend acts as link between Frontend and Knowledge Graph and provides additional features like user management and
roles.
The Backend is a Flask application that uses
rdflib to interface with the knowledge graph.
Take a look at the graph-connector API documentation to get insights into how and with what data to use the backend api.