Skip to main content

A python library for retrieving semantic prefix maps

Project description

prefixmaps

A Python library for retrieving semantic prefix maps.

A semantic prefix map will map a a prefix (e.g. skos) to a namespace (e.g http://www.w3.org/2004/02/skos/core#).

This repository and the corresponding library is designed to satisfy the following requirements:

  • generation of prefix maps in headers of RDF documents
  • use in tools that expand CURIEs and short-form identifiers to URIs that can be used as subjects of RDF triples
  • coverage of prefixes from multiple different domains
  • no single authoritative source of either prefixes or prefix-namespace mappings (clash-resilient)
  • preferred semantic namespace is prioritized over web URLs
  • authority preferred prefix is prioritized where possible
  • each individual prefix map is case-insensitive bijective
  • prefix map composition and custom ordering of prefixmaps
  • lightweight / low footprint
  • fast (TODO)
  • network-independence / versioned prefix maps
  • optional ability to retrieve latest from external authority on network

What this is NOT intended for:

  • a general source of metadata about either prefixes or namespaces
  • a mechanism for resolving identifiers to web URLs for humans to find information

Installation

pip install prefixmaps

Usage

To use in combination with curies library:

from prefixmaps import load_converter
from curies import Converter

converter: Converter = load_converter(["obo", "bioregistry.upper", "linked_data", "prefixcc"])

>>> converter.expand("CHEBI:1")
'http://purl.obolibrary.org/obo/CHEBI_1'
>>> converter.expand("GEO:1")
'http://purl.obolibrary.org/obo/GEO_1'
>>> converter.expand("owl:Class")
'http://www.w3.org/2002/07/owl#Class'
>>> converter.expand("FlyBase:FBgn123")
'http://identifiers.org/fb/FBgn123'

Alternate orderings / clash resilience

  • prefix.cc uses the prefix geo for geosparql http://www.opengis.net/ont/geosparql#
  • OBO uses prefix GEO for the Geographical Entity Ontology, expanding to http://purl.obolibrary.org/obo/GEO_
  • the Bioregistry uses the prefix geo for NCBI GEO, and "re-mints" a geogeo prefix for the OBO ontology

If we prioritize prefix.cc the OBO prefix is ignored:

converter = load_converter(["prefixcc", "obo"])

>>> converter.expand("GEO:1")
>>> converter.expand("geo:1")
'http://www.opengis.net/ont/geosparql#1'

Even though prefix expansion is case-sensitive, we intentionally block conflicts that differ only in case.

If we push bioregistry at the start of the list then GEOGEO can be used as the prefix for the OBO ontology:

converter = load_converter(["bioregistry", "prefixcc", "obo"])

>>> converter.expand("geo:1")
'http://identifiers.org/geo/1'
>>> converter.expand("GEO:1")
>>> converter.expand("GEOGEO:1")
'http://purl.obolibrary.org/obo/GEO_1'

Note that from the OBO perspective, GEOGEO is non-canonical.

We get similar results using the upper-normalized variant of bioregistry:

converter = load_converter(["bioregistry.upper", "prefixcc", "obo"])

>>> converter.expand("GEO:1")
'http://identifiers.org/geo/1'
>>> converter.expand("geo:1")
>>> converter.expand("GEOGEO:1")
'http://purl.obolibrary.org/obo/GEO_1'

Users of OBO ontologies will want to place OBO at the start of the list:

converter = load_converter(["obo", "bioregistry.upper", "prefixcc"])

>>> converter.expand("geo:1")
>>> converter.expand("GEO:1")
'http://purl.obolibrary.org/obo/GEO_1'
>>> converter.expand("GEOGEO:1")

Note under this ordering there is no prefix for NCBI GEO. This is not a major limitation as there is no canonical semantic rendering of NCBI GEO. This could be added in future with a unique OBO prefix.

You can use the ready-made "merged" prefix set, which prioritizes OBO:

converter = load_converter("merged")

>>> converter.expand("GEOGEO:1")
>>> converter.expand("GEO:1")
'http://purl.obolibrary.org/obo/GEO_1'
>>> converter.expand("geo:1")

Network independence and requesting latest versions

By default, this will make use of metadata distributed alongside the package. This has certain advantages in terms of reproducibility, but it means if a new ontology or prefix is added to an upstream source you won't see this.

To refresh and use the latest upstream:

converter = load_converter("obo", refresh=True)

This will perform a fetch from http://obofoundry.org/registry/obo_prefixes.ttl

Context Metadata

See contexts.curated.yaml

See the description fields

Repository organization

Data files containing pre-build prefix maps using sources like OBO and Bioregistry are distributed alongside the python

Location:

CSV field descriptions

  1. context: a unique handle for this context. This MUST be the same as the basename of the file
  2. prefix: corresponds to http://www.w3.org/ns/shacl#prefix
  3. namespace: corresponds to http://www.w3.org/ns/shacl#namespace
  4. canonical: true if this satisfies bijectivity

Refreshing the Data

The data can be refreshed in several ways:

  1. Locally, you can use tox with:

    pip install tox tox-poetry
    tox -e refresh
    
  2. Manually running and automatically committing via this GitHub Actions workflow.

  3. Running makefile (warning, this requires some pre-configuration

    make etl
    

TODO: make a github action that auto-releases new versions

Note that PRs should not be made against the individual CSV files. These are generated from upstream sources.

We temporarily house a small number of curated prefixmaps such as linked_data.yaml, with the CSV generated from the YAML.

Our goal is to ultimately cede these to upstream sources.

Requesting new prefixes

This repo is NOT a prefix registry. Its job is simply to aggregate different prefix maps. Request changes upstream.

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

prefixmaps-0.2.6.tar.gz (709.9 kB view details)

Uploaded Source

Built Distribution

prefixmaps-0.2.6-py3-none-any.whl (754.7 kB view details)

Uploaded Python 3

File details

Details for the file prefixmaps-0.2.6.tar.gz.

File metadata

  • Download URL: prefixmaps-0.2.6.tar.gz
  • Upload date:
  • Size: 709.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.8.20

File hashes

Hashes for prefixmaps-0.2.6.tar.gz
Algorithm Hash digest
SHA256 7421e1244eea610217fa1ba96c9aebd64e8162a930dc0626207cd8bf62ecf4b9
MD5 8ec897710f87653d6c9cbfaf625b285b
BLAKE2b-256 4dcff588bcdfd2c841839b9d59ce219a46695da56aa2805faff937bbafb9ee2b

See more details on using hashes here.

File details

Details for the file prefixmaps-0.2.6-py3-none-any.whl.

File metadata

  • Download URL: prefixmaps-0.2.6-py3-none-any.whl
  • Upload date:
  • Size: 754.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.8.20

File hashes

Hashes for prefixmaps-0.2.6-py3-none-any.whl
Algorithm Hash digest
SHA256 f6cef28a7320fc6337cf411be212948ce570333a0ce958940ef684c7fb192a62
MD5 9d35daf2ca876f35344b8b092a6d01be
BLAKE2b-256 89b22b2153173f2819e3d7d1949918612981bc6bd895b75ffa392d63d115f327

See more details on using hashes here.

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