Skip to content

Latest commit

 

History

History
83 lines (58 loc) · 2.74 KB

README.md

File metadata and controls

83 lines (58 loc) · 2.74 KB
Raw

SPARQL Sleuth

This is a client-side web application which allows the user to discover information from a SPARQL endpoint by interacting with a representation of the underlying knowledge graph.

Screenshot

screenshot

Demo: https://partygate-web-prod-wrennwfy6q-ew.a.run.app/graph?node=http:%2F%2Fe623fd-partygate%2Fdec-top-aide-leaving-do&relationships=both

How it works

This is a web application built using Angular. Building the application results in a set of resources which should be served from a web server.

The application is self-contained once built with the exception that it needs a SPARQL endpoint at the /sparql relative URL. This is not provided by the Angular application, but should be provided by the web service on which the application is hosted.

One option is to get the web server to reverse-proxy /sparql to a real SPARQL service elsewhere on the internet.

The application works by POSTing SPARQL queries to the endpoint.

Web proxy

There is a bundled web proxy/serve.go written in Golang, which can be used to redirect /sparql to another host.

See also

I have a SPARQL service at https://github.com/cybermaggedon/sparql-service which can be used with simple RDF datasets, maybe not good for larger datasets or scalable services beyond 1 billion edges.

Deployment

Container

Easiest way to deploy is to use the containers I have built: docker.io/cybermaggedon/sparql-sleuth which bundles both the built Angular application plus the web proxy. The only thing you need to do is tell the application where the SPARQL endpoint is, and you do this by setting the container command line. It should take the form:

  /usr/local/bin/serve 0:8080 SPARQL_HOST:PORT SCHEME .

but replace SPARQL_HOST with the hostname of your SPARQL service, PORT with the SPARQL endpoint's port number, and replace SCHEME with either http or https depending on which the SPARQL endpoint uses. If it's not set up with TLS, use http.

For a SPARQL service, my repo https://github.com/cybermaggedon/sparql-service can be easily used for small RDF datasets which are built into a sqlite database.

Alternatively

You need to:

  • Build the Angular application with ng build -c production
  • The resultant web application is built to dist/sparql-sleuth, and you need to put that on your web server.
  • Redirect /sparql on your web server to the SPARQL endpoint.

Development

You need to run:

  • ng serve to run the SPARQL application on port 4200.
  • Some sort of proxy. I use the code in proxy/serve-local.go to redirect to the SPARQL endpoint and Angular service.
  • A SPARQL service.

e.g. to list on port 8080, redirect to a SPARQL endpoint on http://localhost:8089:

./serve-local 0:8080 localhost:8089 http .