Loc-I Data Profile - Specification

URI
https://linked.data.gov.au/def/loci-dp/spec
Title
Loc-I Data Profile - Specification Document
Definition
This document specifies the Loc-I Data Profile. It is to be used to inform people about the requirements that need to be met by data claiming to conform to the profile.
Created
2021-04-01
Modified
2021-04-03
Creator
SURROUND Australia Pty Ltd
Publisher
Geoscience Australia
Further metadata
This specification is part of the Loc-I Data Profile. See that profile's main document for License & Rights information and other metadata not given here.
Profile URI
https://linked.data.gov.au/def/loci-dp

Abstract

This is the specification document of the Loc-I Data Profile which constrains the Loc-I Ontology by placing requirements on the classes defined in that ontology in terms of what properties instances of each class must have. The contraints are derived from OWL restrictions in the ontology be also slightly extended to require basic annotation properties too (title, descriptions etc.).

This specification is not to be used for testing conformance of RDF resources to this profile. That role belongs to the multiple validation resource within this profile.

For the list of all resources within this profile, see the profile definition:

Namespaces

This document refers to elements of various ontologies by short codes using namespace prefixes. The prefixes and their corresponding namespaces' URIs are:

dcterms
http://purl.org/dc/terms/
prof
http://www.w3.org/ns/dx/prof/
prov
http://www.w3.org/ns/prov#
sdo
https://schema.org/
skos
http://www.w3.org/2004/02/skos/core#
rdfs
http://www.w3.org/2000/01/rdf-schema#

1. Introduction

The Location Lndex (Loc-I) Project contains complex data and this profile specifies what properties classes of object within that project must exhibit. The properties are defined here in English and individual requirements are given identifiers which machine-readable versions of these requirments quote to enable validation to be traced back to them.

Formally, this document specifies a profile of the Loc-I Ontology and by profile, the definition of from the Profiles Vocabularyref is used. A profile is:

A specification that constrains, extends, combines, or provides guidance or explanation about the usage of other specifications.

Here, the other specification being profiled is the Loc-I Ontology.

In the next section, this document describes how Loc-I Ontology class instances must be presented in data to ensure it conforms to this profile.

This specification's rules/requriements - are numbered and indicated in red text.

2. Requirements

This section lists requirements for Loc-I Ontology class instances.

2.1 Dataset

D-title: Each Dataset MUST have one and only one English title which is an English text literal, indicated using the skos:prefLabel property

skos:prefLabel is to be used rather than rdfs:label or dcterms:title or other, similar, titling properties.

D-defn: Each Dataset MUST have one and only one English definition which is an English text literal, indicated using the skos:definition property

skos:definition is to be used rather than rdfs:comment or dcterms:description or other, similar, description properties.

D-id: Each Dataset MUST have one and only one identifier, an xsd:token, indicated using the dcterms:identifier property

This identifier MUST be unique among Loc-I Datasets. This uniqueness is not testable in SHACL

D-created: Each Dataset MUST have one and only one created date, an xsd:date, xsd:dateTime or xsd:dateTimeStamp value, indicated using the dcterms:created property

D-modified: Each Dataset MUST have one and only one modified date, an xsd:date, xsd:dateTime or xsd:dateTimeStamp value, indicated using the dcterms:modified property

D-creator: Each Dataset MUST indicate one or more creator agents with the dcterms:publisher property, typed as an sdo:Person, sdo:Organization or sdo:GovernmentOrganization

D-publisher: Each Dataset MUST indicate at least one publisher agent with the dcterms:publisher property, typed as an sdo:Organization or sdo:GovernmentOrganization, that is registered by the AGLDWG within their Organisations register

D-pbundle: Each Dataset MUST indicate that it is within a prof:Bundle by an instance of one being indicated with the prov:asInBundle property

D-derived: If a Dataset is derived from anotehr Dataset, it SHOULD indicate that that other dataset, a general DCAT Dataset instnace, with a prov:wasDerivedFrom property

2.2 Feature Collection

Loc-I matches the OGC APIref specification's method of arranging multiple Feature instances within a Dataset by grouping them into FeatureCollection. FeatureCollection instances are not spatial object themselves but are administrative (managed) collection (of Feature instances). They just require basic metadata, no spatial properties.

FC-title: Each Feature Collection MUST have one and only one English title which is an English text literal, indicated using the skos:prefLabel property

FC-defn: Each Feature Collection MUST have one and only one English definition which is an English text literal, indicated using the skos:definition property

FC-id: Each Feature Collection MUST have one and only one identifier, an xsd:token, indicated using the dcterms:identifier property. Note: this identifier must be unique within the Dataset it is part of. This uniqueness is not testable in SHACL.

FC-part: Each Feature Collection MUST indicate that it is part of one and only one dcat:Dataset with use of the dcterms:isPartOf property

2.3 Feature

Features are the code things within Loc-I data. The are the spatial objects we care about and to which properties of interest are attached. Loc-I requirements for Feature isntances are that they have geometries - without which they can't be used as spatial objects - and are part of FeatureCollection instances. Each Feature instance must have an ID that is unique within a FeatureCollection so it can be used by systems taht cannot handle IRIs, such as OGC APIs, to refer to individuals.

F-part: Each Feature MUST indicate that it is part of at least geo:FeatureCollection with use of the dcterms:isPartOf property

FC-id: Each Feature MUST have one and only one identifier, an xsd:token, indicated using the dcterms:identifier property. Note: this identifier must be unique within a Feature Collection. This uniqueness is not testable in SHACL

FC-geometry: Each Feature Collection MUST have at least one geometry indicated using the geo:hasGeometry property

2.4 Linkset

Loc-I Dataset instances may be joined by Linkset instances. A Linkset is a specialised form of a Dataset specifically created to join other Dataset instances and, as a result, has some mandatory properties to ensure joining is correctly specified.

Note that because Linkset instances are also Dataset instances, all the Dataset rule apply to Linkset instances in addition to thiese specialised requirements.

L-m: Each Linkset MUST indicate at least one rdf:Statement by an inverse linking from that instance to the Linkset via the dcterms:isPartOf property

L-st: Each Linkset MUST indicate one an only one subjects target Dataset with the property void:subjectsTarget

L-ot: Each Linkset MUST indicate one an only one objects target Dataset with the property void:objectsTarget

2.5 Statement

The graph elements within Linkset instances that indicate Feature-level links between Dataset instances are Statement instances which are specialised forms of the standard rdf:Statement class.

LS-subject: Each Statement MUST indicate one an only one subject with the propertyrdf:subject

LS-predicate: Each Statement MUST indicate one an only one predicate with the propertyrdf:predicate

LS-object: Each Statement MUST indicate one an only one object with the propertyrdf:object

L-part: Each Statement MUST indicate that it is part of one and only onevoid:Linkset with use of thedcterms:isPartOf property

LS-method: Each Statement MUST indicate that it was generated by one and only one method (aprov:Plan) with use of theloci:hadGenerationMethod property

2.6 Bundle

Provenance for Loc-I datasets must be recorded using PROV-O, the Provenance Ontologyref. The provenance statements are collected in prov:Bundle objects. EAch Bundle instance must include certain elements that allow users to understand how the Dataset that the Bundle is for was produced.

B-mainactivity: Each Bundle MUST contain at least one prov:Activity indicated by a dcterms:hasPart property which must, in turn, indicate at least one prov:Entity by use of prov:generated

3. References

OGC-API
Clemens Portele, Panagiotis (Peter) A. Vretanos, Charles Heazel (eds.). OGC API - Features - Part 1: Core. 14 October 2019. Open Geospatial Consortium standard. URL: http://www.opengis.net/doc/IS/ogcapi-features-1/1.0
PROF
Rob Atkinson; Nicholas J. Car (eds.). The Profiles Vocabulary. 18 December 2019. W3C Working Group Note. URL: https://www.w3.org/TR/dx-prof/
PROV
Lebo, Timothy, Satya Sahoo, and Deborah McGuinness. PROV-O: The PROV Ontology. W3C Recommendation. W3C Provenance Working Group, 2013. URL: http://www.w3.org/TR/prov-o/.