Ldtools 0.6.0 (beta)

A lightweight ORM for Linked Data

Installation

Just use pip:

pip install Ldtools

This will install all dependencies (argparse, rdflib) needed and provide the command line utility "ldtools".

Alternatively, do a git clone and execute python setup.py install/develop.

How to use it?

Via the commandline, all information within a Linked Data resource can be retrieved by executing:

ldtools http://dbpedia.org/resource/Karlsruhe

Further options can be utilised to influence whether the URIs that are discovered within the origin should be followed and how deep. Try:

ldtools --help

for more usage information.

Alternatively, the python console can be used:

First, we create an Origin object:

import pprint, rdflib
from ldtools.resource import Resource
from ldtools.origin import Origin

uri = "http://dbpedia.org/resource/Karlsruhe"
origin, created = Origin.objects.get_or_create(uri)

Process it, hopefully discovering more Origins in there (rdfs:seeAlso, owl:sameAs...):

origin.GET(only_follow_uris=[rdflib.OWL.sameAs,rdflib.RDFS.seeAlso])

If everything went well, there now is a Resource object for our uri:

resource, created = Resource.objects.get_or_create(uri, origin=origin)
pprint.pprint(resource.__dict__)

Process all the other Origins we know about:

Origin.objects.GET_all()

Result: 5 URIs crawled and 500 Resources discovered and processed.

Why?

• The Semantic Web is out there and there is really not enough tools yet to work with Linked Data
• SPARQL is not needed to get the RAW data from resources, this library demonstrates that. Just the basic Linked Data Stack: URIs, Content Negotiation, RDF needed
• ldtools intends to make it easy to handle the data you get from an URI and to follow links you discover
• Based on that, you can modify your objects and PUT them back to their origin

Tests

To run the tests, install spec and/or nose and run nose:

pip install spec unittest2 coverage
nosetests --with-coverage --cover-package=ldtools
nosetests --with-specplugin

Build Status

Contributions/Credits

Feel free to submit ideas and bugs to http://github.com/dmr/Ldtools/issues, I'll be happy to accept pull requests for new features.

Thank you Travis CI for running the tests :)

Thanks to Django, Flask, peewee and sentry for inspiration regarding model structure!