RelFinder API

Relfinder is a tool for calculating and visualising paths between entities (nodes) in an RDF Knowledge Graph (or SPARQL endpoint). It is based on the original Relfinder tool developed by Philipp Heim, Steffen Lohmann, Timo Stegemann and others. For more information about the tool and its features, please see: http://www.visualdataweb.org/relfinder.php.

The Relfinder source code in this repository is a Python (Flask) porting of the backend of the original Relfinder tool which was developed in PHP. We have also developed a front-end for this Flask API here using Vue.js: http://github.com/nwa-catch-me-if-you-can-project/relfinder-frontend

Motivation for development

The original Relfinder had a front-end developed with Adobe Flash / Flex and since this technology has been discontinued, there was a need to preserve the useful functionality of this tool for Semantic Web research and the community. Hence the motivation for porting the tool to Python.

Development setup

To run the app in development mode run the following commands. Before running the app create a .env file (the name of this file must strictly be .env) as specified in the production setup section

export FLASK_APP=api
export FLASK_ENV=development
flask run

Production setup

Before running the application create a .env file (the name of the file must strictly be .env) using the template in example.env. The template will look like this:

SPARQL_ENDPOINT=*** sparql endpoint url ***
SPARQL_USERNAME=*** sparql endpoint username ***
SPARQL_PASSWORD=*** sparql endpoint pwd ***
API_KEY=*** arbitrary api key (this is just to "pair" with the front-end which should specify the same api key in its .env file) ***
DEBUG=TRUE
ONTOLOGY_PREFIX=*** namespace for main ontology or vocabulary used in the graph ***

Note: the SPARQL_ENDPOINT address should be an external address (e.g. http://dbpedia.org/sparql). If you would like to test locally using localhost or another container, you need to ensure the address is visible to this container.

The API_KEY variable is arbitrary to connect the backend to the frontend. We recommend generating such a key with a tool such as this. If you would like to use the RelFinder frontend as well, you also need to configure VUE_APP_API_KEY to have the same value as API_KEY in the frontend's .env file.

Once you have created the .env file, build the docker image with:

docker build . -t relfinder-api:0.0.1

To run the container execute

docker run --rm -p 5000:5000 relfinder-api:0.0.1

License and contributions

The Relfinder API is developed and copyrighted by Kody Moodley and Walter Simoncini, and released under a dual license (see LICENSE.md file in this repository) Contributions and bug reports are helpful and welcome.