Skip to content

typingeek/hydra-py

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hydra library for Python

The primary goal is to provide a lib for easily writing Hydra-enabled clients [1].

A secondary goal is to provide a client for Triple Patterns Fragments [2], and an RDFlib [3] Store backed on any TPF service.

Installation

To install this library, from the projet directory, type:

pip install .

NB: developers might want to add the -e option to the command line above, so that modifications to the source are automatically taken into account.

Quick start

To create a Hydra-enabled resource, use:

from hydra import Resource, SCHEMA
res = Resource.from_iri(the_iri_of_the_resource)

If the resource has an API documentation associated with it, it will be available as an attribute. The API documentation provides access to the supported class, their supported properties and operations.

print("Api documentation:")
for supcls in res.api_documentation.supported_classes:
    print("  %s" % supcls.identifier)
    for supop in supcls.supported_operations:
        print("    %s" % supop.identifier)

Alternatively, you can query the resource directly for available operations. For example, the following searches for a suitable operation for creating a new event, and performs it.

create_event = res.find_suitable_operation(SCHEMA.AddAction, SCHEMA.Event)
resp, body = create_event({
    "@context": "http://schema.org/",
    "@type": "http://schema.org/Event",
    "name": "Halloween",
    "description": "This is halloween, this is halloween",
    "startDate": "2015-10-31T00:00:00Z",
    "endDate": "2015-10-31T23:59:59Z",
})
assert resp.status == 201, "%s %s" % (resp.status, resp.reason)
new_event = Resource.from_iri(resp['location'])

And you can go on with the new event you just created...

Triple Pattern Fragments

The hydra.tpf module implements of Triple Pattern Fragments specification [2]. In particular, it provides an implementation of Store, so that TPF services can be used transparently:

import hydra.tpf # ensures that the TPFStore plugin is registered
from rdflib import Graph

g = Graph('TPFStore')
g.open('http://data.linkeddatafragments.org/dbpedia2014')

results = g.query("SELECT DISTINCT ?cls { [ a ?cls ] } LIMIT 10")

Note however that this is experimental at the moment...

References

[1]http://www.hydra-cg.com/
[2](1, 2) http://www.hydra-cg.com/spec/latest/triple-pattern-fragments/
[3]https://rdflib.readthedocs.org/

About

Hydra implementation of Python

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%