sparqlowl
best practice regarding separation of ontology and data for SPARQL queries?
I'm working on an ontology and a linked open dataset that uses it. I am using Protégé 5.5.0 on a Mac. I have tried running SPARQL queries in Protégé and I have been using stain/jena-fuseki on docker as well.
I would like to separate the ontology (contained in vocab.owl) from the data (data.ttl)
In vocab.owl, I have created classes Item and Scale. An Item has a Scale, one of 1_One, 2_Two, and 3_Three. Each Scale has a value and a label.
vocab.owl
@prefix tst: <http://www2.cs.uregina.ca/~hepting/vocab.owl#> .
@prefix : <http://www2.cs.uregina.ca/~hepting/data.ttl#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix xml: <http://www.w3.org/XML/1998/namespace> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@base <http://www2.cs.uregina.ca/~hepting/vocab.owl> .
<http://www2.cs.uregina.ca/~hepting/vocab> rdf:type owl:Ontology .
#################################################################
# Object Properties
#################################################################
tst:hasScale rdf:type owl:ObjectProperty .
#################################################################
# Data properties
#################################################################
tst:hasScale rdf:type owl:DatatypeProperty .
tst:hasScaleLabel rdf:type owl:DatatypeProperty ;
rdfs:subPropertyOf tst:hasScale .
tst:hasScaleValue rdf:type owl:DatatypeProperty ;
rdfs:subPropertyOf tst:hasScale .
#################################################################
# Classes
#################################################################
tst:Item rdf:type owl:Class ;
rdfs:subClassOf [ rdf:type owl:Restriction ;
owl:onProperty tst:hasScale ;
owl:someValuesFrom tst:Scale
] .
tst:Scale rdf:type owl:Class ;
owl:equivalentClass [ rdf:type owl:Class ;
owl:oneOf ( tst:1_One
tst:2_Two
tst:3_Three
)
] .
#################################################################
# Named Individuals
#################################################################
tst:1_One a owl:NamedIndividual , tst:Scale ;
tst:hasScaleValue "1"^^xsd:positiveInteger ;
tst:hasScaleLabel "One"^^xsd:string .
tst:2_Two a owl:NamedIndividual , tst:Scale ;
tst:hasScaleValue "2"^^xsd:positiveInteger ;
tst:hasScaleLabel "Two"^^xsd:string .
tst:3_Three a owl:NamedIndividual , tst:Scale ;
tst:hasScaleValue "3"^^xsd:positiveInteger ;
tst:hasScaleLabel "Three"^^xsd:string .
In data.ttl, I have created 7 instances of Item, that each have an instance of Scale.
data.ttl:
@prefix tst: <http://www2.cs.uregina.ca/~hepting/vocab.owl#> .
@prefix : <http://www2.cs.uregina.ca/~hepting/data.ttl#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix xml: <http://www.w3.org/XML/1998/namespace> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
:Item1 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:1_One .
:Item2 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:2_Two .
:Item3 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:3_Three .
:Item4 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:3_Three .
:Item5 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:3_Three .
:Item6 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:3_Three .
:Item7 a owl:NamedIndividual , tst:Item ;
tst:hasScale tst:3_Three .
I want to run a query that associates Items with ScaleValues.
query:
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX tst: <http://www2.cs.uregina.ca/~hepting/vocab.owl#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX : <http://www2.cs.uregina.ca/~hepting/data.ttl#>
SELECT * WHERE
{
?item tst:hasScale ?scale .
?scale tst:hasScaleValue ?value .
}
ORDER BY ?scale
In docker stain/jena-fuseki, I load data.ttl into a dataset (21 triples) and my query above has no errors but returns no results. (The results are similar if I use the SPARQL Query tab to execute my query in Protégé with data.ttl loaded)
If I edit my dataset in docker stain/jena-fuseki to add the following:
tst:1_One a owl:NamedIndividual , tst:Scale ;
tst:hasScaleValue "1"^^xsd:positiveInteger ;
tst:hasScaleLabel "One"^^xsd:string .
tst:2_Two a owl:NamedIndividual , tst:Scale ;
tst:hasScaleValue "2"^^xsd:positiveInteger ;
tst:hasScaleLabel "Two"^^xsd:string .
tst:3_Three a owl:NamedIndividual , tst:Scale ;
tst:hasScaleValue "3"^^xsd:positiveInteger ;
tst:hasScaleLabel "Three"^^xsd:string .
to give 33 triples, I get the following raw results for my query -- 1 entry for each of the 7 items in data.ttl (the desired results):
{
"head": {
"vars": [ "item" , "scale" , "value" ]
} ,
"results": {
"bindings": [
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item1" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#1_One" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "1" }
} ,
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item2" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#2_Two" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "2" }
} ,
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item3" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#3_Three" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "3" }
} ,
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item4" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#3_Three" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "3" }
} ,
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item5" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#3_Three" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "3" }
} ,
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item6" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#3_Three" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "3" }
} ,
{
"item": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/data.ttl#Item7" } ,
"scale": { "type": "uri" , "value": "http://www2.cs.uregina.ca/~hepting/vocab.owl#3_Three" } ,
"value": { "type": "literal" , "datatype": "http://www.w3.org/2001/XMLSchema#positiveInteger" , "value": "3" }
}
]
}
}
Returning to the 21 triples in the data.ttl file, I tried the following query (with the addition of FROMs) that returned no data.
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX tst: <http://www2.cs.uregina.ca/~hepting/vocab.owl#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX : <http://www2.cs.uregina.ca/~hepting/data.ttl#>
SELECT *
FROM <http://www2.cs.uregina.ca/~hepting/vocab.owl>
FROM <http://www2.cs.uregina.ca/~hepting/data.ttl>
WHERE
{
?s ?p ?o
}
Is there a way to get the desired results while maintaining the separation between vocab.owl and data.ttl?
Solution
You have to upload both Turtle files (the data and the ontology) to Fuseki.
A potential pitfall: Fuseki doesn’t allow uploading Turtle files with the extension .owl. Unfortunately, it fails somewhat silently, by coloring the progress bar in red, and printing "Triples uploaded: 0".
After renaming the ontology to vocab.ttl, and uploading it, the first query from your question works as expected:
- Math.Sin() gives incorrect value
- How to run my python script when the sunOS is start booting
- Express-session: not resetting cookie expiration on each request
- Getting a stack overflow exception when normalizing a vector
- Edit default summary function in R gives error for multiple variables
- What was a For loop? Why isn't it needed in R?
- How to use download button in shiny and save results in various formats (csv, texte, pdf, spss...)?
- Why are there two assignment operators, `<-` and `->` in R?
- lm()$assign: what is it?
- How to get the value of list(...) in R and S functions
- Design matrix for MLM from library(lme4) with fixed and random effects
- how to generate elements not included in my sample
- Create a matrix with gradually changing values without a for loop
- Emacs ESS and S-plus ( S+ ) 8.1 compatability
- How to lag date-index in a time-series in R?
- Nonlinear regression in R / S
- Calling R from S-Plus?