Python4Spongers
RDF "Spongers" are a powerful middleware architecture, developed by OpenLink Software (the makers of Virtuoso), for creating RDF rich meta-data on demand.
The key idea is that the middleware consults public APIs or other data sources for collating relevant RDF meta-data for a given URI.
Unfortunately, the development of such sponger components is still difficult for many programmers. On this page, I propose a simple skeleton for coding the core transformation in Python.
This still requires a wrapper so that the code can be used in a Virtuoso environment, but that should be doable.
If you have any questions or suggestions, please contact me at mheppATcomputerDOTorg.
Example:
- !/usr/bin/env python
- encoding: utf-8
""" py4spongers.py
Example of how the principle of OpenLink "Sponger" technology can be implemented in Python
Created by Martin Hepp on 2010-03-22. http://www.heppnetz.de/
This software is free software under the LPGL.
"""
import re from rdflib import *
def rdf4uri(uri="", base_uri="http://example.com/uriburner/"):
"""
This method returns available RDF meta-data for the Web page identified by `` as a string containing RDF/XML.
Input Parameters:
uri : URI of the page
base_uri : Base URI to be used for the RDF model
Output Parameter
a string containing RDF/XML
"""
# Step 1: Fetch entity identifier from URI
# Amazon Example:
# http://www.amazon.com/Apple-touch-Generation-NEWEST-MODEL/dp/B002M3SOBU/ref=sr_1_9?ie=UTF8&s=electronics&qid=1269264339&sr=8-9
# http://www.amazon.com/Apple-touch-Generation-NEWEST-MODEL/dp/B002M3SOBU/
#
# We use a simple regex to extract the ID from the URI (needs to be adapted per each sponger)
p = re.compile(r".*/dp/(\w*)/.*")
m = p.match(uri)
identifier = m.group(1)
# Step 2: Fetch meta-data for that data entity, e.g. via AMAZON API
# contact API --> omitted in this example
# In this example, we simply return static data
# Step 3: Compile RDF Graph
NS = Namespace(base_uri)
GR = Namespace('http://purl.org/goodrelations/v1#')
RDFS = Namespace("http://www.w3.org/2000/01/rdf-schema#")
RDF = Namespace("http://www.w3.org/1999/02/22-rdf-syntax-ns#")
g = ConjunctiveGraph()
# static dummy data, to be replaced by real content from API
g.add((NS[identifier+'#Product'],RDF['type'],GR['ProductOrServicesSomeInstancesPlaceholder']))
g.add((NS[identifier+'#Product'],RDFS['label'],Literal('SampleProduct')))
# Step 4: Return Graph as RDF/XML
return g.serialize()
if name == 'main':
rdf_xml = rdf4uri(uri='http://www.amazon.com/Apple-touch-Generation-NEWEST-MODEL/dp/B002M3SOBU/')
print rdf_xml
In the current form, it will return a static pattern for each valid Amazon product URI:
``
``
``
``