Skip to main content

rdflib stores based on pyoxigraph

Project description

Oxrdflib

PyPI PyPI - Implementation PyPI - Python Version actions status Gitter

Oxrdflib provides an rdflib store based on pyoxigraph. This store is named "Oxigraph".

This store can be used as drop-in replacement of the rdflib default one. It support context but not formulas. Transaction support is not implemented yet.

SPARQL query evaluation is done by pyoxigraph instead of rdflib if the Oxigraph store is used. SPARQL update evaluation is still done using rdflib because of a limitation in rdflib context management.

Oxrdflib is available on Pypi and installable with:

pip install oxrdflib

The oxrdflib store is automatically registered as an rdflib store plugin by setuptools.

Warning: Oxigraph is not stable yet and its storage format might change in the future. To migrate to future version you might have to dump and load the store content. However, Oxigraph should be in a good enough shape to power most of use cases if you are not afraid of down time and data loss.

API

To create a rdflib graph using the Oxigraph store use

rdflib.Graph(store="Oxigraph")

instead of the usual

rdflib.Graph()

Similarly, to get a conjunctive graph, use

rdflib.ConjunctiveGraph(store="Oxigraph")

instead of the usual

rdflib.ConjunctiveGraph()

and to get a dataset, use

rdflib.Dataset(store="Oxigraph")

instead of the usual

rdflib.Dataset()

If you want to get the store data persisted on disk, use the open method on the Graph object (or ConjunctiveGraph or Dataset) with the directory where data should be persisted. For example:

graph = rdflib.Graph(store="Oxigraph", identifier="http://example.com") # without identifier, some blank node will be used
graph.open("test_dir")

The store is closed with the close() method or automatically when Python garbage collector collects the store object.

If the open method is not called Oxigraph will automatically use a ramdisk on Linux and a temporary file in the other operating systems.

To do anything else, use the usual rdflib python API.

It is also possible to directly inject a pyoxigraph Store object directly into an Oxrdflib store:

graph = rdflib.Graph(store=oxrdflib.OxigraphStore(store=pyoxigraph.Store("test_dir")))

This might be handy to e.g. open the database as read-only:

graph = rdflib.Graph(store=oxrdflib.OxigraphStore(store=pyoxigraph.Store.read_only("test_dir")))

Differences with rdflib default store

  • relative IRIs are not supported by Oxigraph.
  • Just like the SPARQLStore, Oxigraph joins the initBindings parameter of the query method after the query has been evaluated, instead of injecting them at the beginning of the query.
  • IRI prefixes set using the Graph bind method are not persisted on disk but kept in memory. They should be added again each time the store is opened.

Migration guide

From 0.2 to 0.3

  • The 0.2 stores named "OxSled" and "OxMemory" have been merged into the "Oxigraph" store.
  • The on-disk storage system provided by "OxSled" has been dropped and replaced by a new storage system based on RocksDB. To migrate you need to first dump your data in RDF using oxrdflib 0.2 and the serialize method, then upgrade to oxrdflib 0.3, and finally reload the data using the parse method.

Development

To run the test do first pip install -e . to register the stores in rdflib plugin registry. Then, cd tests && python -m unittest should run the tests.

The code is automatically formatted using black. A pre-commit configuration is provided. Run pip install pre-commit && pre-commit install to install pre-commit as a git pre-commit hook in your clone.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

oxrdflib-0.3.7.tar.gz (13.6 kB view hashes)

Uploaded Source

Built Distribution

oxrdflib-0.3.7-py3-none-any.whl (7.2 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page