Part 4: Working with RDF
Previously we saw how to do basic validation of data using JSON-Schema.
This section demonstrates how to work with LinkML in conjunction with Linked Data/ RDF. If this is not of interest you can skip to the next section. However, even if this is the case you may wish to revisit this section. LinkML is intended to make it easy to get the benefits of Linked Data, while staying simple and working within a stack many developers are familiar with.
And even if you aren’t using RDF yourself, declaring URIs for your schema elements can help make your data FAIR, and in particular can serve as hooks to interoperate data!
Example schema
Let’s start with the schema we developed in the previous section:
personinfo.yaml:
id: https://w3id.org/linkml/examples/personinfo
name: personinfo
prefixes:
linkml: https://w3id.org/linkml/
imports:
- linkml:types
default_range: string
classes:
Person:
attributes:
id:
identifier: true
full_name:
required: true
description:
name of the person
aliases:
multivalued: true
description:
other names for the person
phone:
pattern: "^[\\d\\(\\)\\-]+$"
age:
range: integer
minimum_value: 0
maximum_value: 200
Container:
attributes:
persons:
multivalued: true
inlined_as_list: true
range: Person
As well as a collection of person records
data.yaml:
persons:
- id: ORCID:1234
full_name: Clark Kent
age: 33
phone: 555-555-5555
- id: ORCID:4567
full_name: Lois Lane
age: 34
We can use the linkml conversion library to translate this to RDF (Turtle syntax default)
linkml-convert -s personinfo.yaml -t rdf data.yaml
Outputs:
@prefix ns1: <https://w3id.org/linkml/examples/personinfo/> .
<ORCID:1234> ns1:age "33" ;
ns1:full_name "Clark Kent" ;
ns1:phone "555-555-5555" .
<ORCID:4567> ns1:age "34" ;
ns1:full_name "Lois Lane" .
[] a ns1:dict ;
ns1:persons <ORCID:1234>,
<ORCID:4567> .
This is a start, but it is not ideal - there are existing vocabularies such as schema.org we could be reusing.
There are also some serious problems - there URIs for our two people are not syntactically valid. We need to provide additional information
Adding URIs to our schema
Let’s enhance our schema, using two schema slots:
class_uri: to declare a URI/IRI for a class
slot_uri: the same thing for a slot
In both cases, we provide the value as a CURIE, and include a prefixes map that maps CURIEs to URIs.
personinfo-semantic.yaml:
id: https://w3id.org/linkml/examples/personinfo
name: personinfo
prefixes: ## Note are adding 3 new ones here
linkml: https://w3id.org/linkml/
schema: http://schema.org/
personinfo: https://w3id.org/linkml/examples/personinfo/
ORCID: https://orcid.org/
imports:
- linkml:types
default_range: string
classes:
Person:
class_uri: schema:Person ## reuse schema.org vocabulary
attributes:
id:
identifier: true
full_name:
required: true
description:
name of the person
slot_uri: schema:name ## reuse schema.org vocabulary
aliases:
multivalued: true
description:
other names for the person
phone:
pattern: "^[\\d\\(\\)\\-]+$"
slot_uri: schema:telephone ## reuse schema.org vocabulary
age:
range: integer
minimum_value: 0
maximum_value: 200
id_prefixes:
- ORCID
Container:
attributes:
persons:
multivalued: true
inlined_as_list: true
range: Person
Now let’s try converting the same YAML/JSON using the enhanced schema
linkml-convert -s personinfo-semantic.yaml -t rdf data.yaml
Outputs:
@prefix ns1: <https://w3id.org/linkml/examples/personinfo/> .
@prefix ns2: <http://schema.org/> .
<https://orcid.org/1234> ns2:name "Clark Kent" ;
ns2:telephone "555-555-5555" ;
ns1:age "33" .
<https://orcid.org/4567> ns2:name "Lois Lane" ;
ns1:age "34" .
[] a ns1:dict ;
ns1:persons <https://orcid.org/1234>,
<https://orcid.org/4567> .
Note that the prefixes are hidden but the effect is to reuse URIs such as schema:telephone
JSON-LD contexts
Behind the scenes, the linkml-convert tool is using JSON-LD contexts to convert from JSON to RDF. This is in the following parts:
A JSON-LD context is generated from the schema
the native JSON representation is enhanced with this context, to make it JSON-LD
A standard JSON-LD converter (in this case, rdflib) is used to convert to RDF
gen-jsonld-context --no-metadata personinfo-semantic.yaml
Outputs:
{
"@context": {
"ORCID": "https://orcid.org/",
"linkml": "https://w3id.org/linkml/",
"personinfo": "https://w3id.org/linkml/examples/personinfo/",
"schema": "http://schema.org/",
"@vocab": "https://w3id.org/linkml/examples/personinfo/",
"persons": {
"@type": "@id"
},
"age": {
"@type": "xsd:integer"
},
"full_name": {
"@id": "schema:name"
},
"id": "@id",
"phone": {
"@id": "schema:telephone"
},
"Person": {
"@id": "schema:Person"
}
}
}
NOTE: currently you need to declare your own type object and map this to rdf:type
for typing information to be shown
Using Shape Languages
In the previous section we saw how to use JSON-Schema validators. LinkML also allows the use of ShEx (future versions will allow SPARQL)
gen-shex --no-metadata personinfo-semantic.yaml
Outputs:
BASE <https://w3id.org/linkml/examples/personinfo/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX linkml: <https://w3id.org/linkml/>
PREFIX schema: <http://schema.org/>
linkml:String xsd:string
linkml:Integer xsd:integer
linkml:Boolean xsd:boolean
linkml:Float xsd:float
linkml:Double xsd:double
linkml:Decimal xsd:decimal
linkml:Time xsd:dateTime
linkml:Date xsd:date
linkml:Datetime xsd:dateTime
linkml:Uriorcurie IRI
linkml:Uri IRI
linkml:Ncname xsd:string
linkml:Objectidentifier IRI
linkml:Nodeidentifier NONLITERAL
<Container> CLOSED {
( $<Container_tes> <persons> @<Person> * ;
rdf:type [ <Container> ] ?
)
}
<Person> CLOSED {
( $<Person_tes> ( schema:name @linkml:String ;
<aliases> @linkml:String * ;
schema:telephone @linkml:String ? ;
<age> @linkml:Integer ?
) ;
rdf:type [ schema:Person ]
)
}
RDF support is evolving
RDF support in LinkML is evolving, and future versions will have more direct support.
More Info
Working with RDF Data
FAQ:
Generators: