W3C home > Mailing lists > Public > public-rww@w3.org > November 2012

Fwd: Trilby - a PHP app for dynamic publishing of small RDF datasets

From: Melvin Carvalho <melvincarvalho@gmail.com>
Date: Sat, 17 Nov 2012 13:39:15 +0100
Message-ID: <CAKaEYhJGenN_PRkRp_-O5iWyxV4LkS=7=T1r7qMoh5wQ9U75yw@mail.gmail.com>
To: public-rww <public-rww@w3.org>

---------- Forwarded message ----------
From: Keith Alexander <keithalexander@keithalexander.co.uk>
Date: 17 November 2012 13:18
Subject: Trilby - a PHP app for dynamic publishing of small RDF datasets
To: linked-data-api-discuss@googlegroups.com


I've written this thing called Trilby[1], and I thought that some people on
this list might find it interesting.

It's not quite another Linked Data API spec implementation, but it's
heavily influenced by LDA.

The niche Trilby is trying to address is the small dataset with thousands
rather than millions of entities. The kind of dataset you might produce
with Open (Google) Refine and the DERI RDF extension.

For small datasets like that, it can be a lot of bother to set up SPARQL
servers and configure APIs and so on, and I thought it would be fun to try
to make it easy to publish small linked data projects that still have good
features for browsing, faceting and searching.

Trilby is a pure PHP(5.3) app that lets you upload an RDF file, and
publishes it dynamically in a fashion similar to LDA.
It uses a PHP file-based RDF store called Raffles[2], and doesn't need a
database server, or require (or provide) a SPARQL endpoint.

Trilby supports paging and path queries eg:

http://keithalexander.co.uk/ecco-tcp/?rdf:type=bibo:Book&_page=2 (also
supports _pageSize)
http://keithalexander.co.uk/ecco-tcp/?dct:creator/whois:born=1750 (deeper
paths - paths also support wildcards eg: */foaf:made= )
http://keithalexander.co.uk/ecco-tcp/?whois:born;_min=1750 (like LDA 'min-'
and 'max-')
http://keithalexander.co.uk/ecco-tcp/?dct:title;_search=Politics (text

Trilby just uses CURIEs rather than LDA shortnames, and doesn't have the
option of "Simple" formats, or configurable URI paths. It doesn't
permit/require much configuration - just a form you fill in where you give
the dataset a name and a license.

Other features include facets, and VoID metadata.

If anyone tries it, I'd love to hear how you get on, and would appreciate
any bugs filed through github etc



[1] http://kwijibo.github.com/trilby/
[2] http://kwijibo.github.com/Raffles/
Received on Saturday, 17 November 2012 12:39:42 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:40:04 UTC