Requirements
In this section we are going to describe in detail what are the requirements that we make on the RDF datasets that can be indexed by QAnswer.
Label
To be able to answer questions, QAnswer needs to know how to translate a question into a SPARQL query. To do so, QAnswer uses the labels attached to the URIs in the RDF dataset.
A URI will only be used to generate SPARQL queries if the question contains (up to stemming) the literal attached via one of the following properties:
If the literal has a language tag (like @en), it will only be searched in questions of that language. If no language tag is attached the corresponding label will be searched in questions of any language.
For example for the following RDF graph:
PREFIX vsw: <http://vocabulary.semantic-web.at/cocktails>
PREFIX vswo: <http://vocabulary.semantic-web.at/cocktail-ontology>
PREFIX schema: <http://schema.org/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
vsw:2d85fb1b rdfs:label "Margarita" ;
rdfs:label "Upside Down Margarita"@en .
vswo:consists-of rdfs:label "consists of"@en ;
rdfs:label "contains"@en ;
rdfs:label "made up"@en ;
rdfs:label "ingredients"@en .
The URI vsw:2d85fb1b will only be used in a SPARQL query if the question either contains "Margarita" or "Upside Down Margarita".
Moreover "Margarita" will be used for any language while "Upside Down Margarita" only for english questions.
The URI vswo:consists-of will be used to construct SPARQL queries if the question contains "consists of", "contains", "made up", "ingredients" and expressions that are equal to them up to stemming. This is for example the case for the expression "contained" which, up to stemming, is the same as "contains".
Note that in particular for the graph:
vsw:Margarita vswo:consists-of vsw:Cointreau .
it will not be possible to answer any question since no labels are attached.
Even if for humans the name of the URI is meaningfull, according to RDF standard the above graph is equivalent to:
vsw:2d85fb1b vswo:1234 vsw:1439e6c3 .
Reification
In QAnswer we assume that the RDF dataset does not use any form of reification.
To recall, RDF is perfectly suited to represent binary statements like: "Margarita contains Cointreau" which can be represented as the triple (Margarita,contains, Cointreau).
Reified statements are used when there is the need to speak about a binary statement like in: "Margarita contains 50ml of Cointreau". In this case a triple is not enough to represent this piece of information. The Semantic Web Community proposed a series of models to represent these types of information. For a full overview of the presented models we refer to this paper. One of the models is n-ary relations (the reification model used by Wikidata), where the knowledge would be represented as:
vsw:Margarita vswo:consists-of_IN _:b1 .
_:b1 vswo:consists-of_OUT vsw:Cointreau .
vsw:blank vswo:quantity "50 ml" .
Another model is RDF reification which was , where the knowledge would be represented as:
vsw:Statement rdf:type rdf:Statement .
vsw:Statement rdf:subject vsw:Margarita .
vsw:Statement rdf:predicate vswo:consists-of .
vsw:Statement rdf:object vsw:Cointreau .
vsw:Statement vswo:quantity "50 ml" .
QAnswer was not designed to cope with these representations and it is not clear how it behaves when they are indexed in QAnswer.
Now let's see how to upload and index your dataset in QAnswer!