Provide an option for a more focused VoID description computation that would take less time, but focus only on the most important informations

[vemonet]

For most our use-cases where we use the VoID description as a schema for knowing the content of an endpoint we only use propertyPartitions that are linked to subject and object classPartitions, such as shared in #24. Basically we just use the info from the query below, which enables to know the most detailed information about the endpoints: 1000 distinct Pathways isPartOf 500 distinct AnatomicalEntities

PREFIX up: <http://purl.uniprot.org/core/>
PREFIX void: <http://rdfs.org/ns/void#>
PREFIX void-ext: <http://ldf.fi/void-ext#>
SELECT DISTINCT ?subjectClass ?prop ?objectClass ?objectDatatype
WHERE {
  {
    ?cp void:class ?subjectClass ;
        void:propertyPartition ?pp .
    ?pp void:property ?prop .
    OPTIONAL {
        {
            ?pp  void:classPartition [ void:class ?objectClass ] .
        	
        } UNION {
            ?pp void-ext:datatypePartition [ void-ext:datatype ?objectDatatype ] .
        }
    }
  } UNION {
    ?linkset void:subjectsTarget [ void:class ?subjectClass ] ;
      void:linkPredicate ?prop ;
      void:objectsTarget [ void:class ?objectClass ] .
  }
}

So there are a lot of computation that are done that we end up never using, e.g. from the Wikipathways VoID:

<https://rdf.wikipathway.org/.well-known/void#rdf.wikipathways.org!1cddc767!DataNoderdf.wikipathways.org!1d576fb0!groupRef>
  a void:Dataset;
  void:property <http://vocabularies.wikipathways.org/gpml#groupRef>;
  void_ext:datatypePartition <https://rdf.wikipathway.org/.well-known/void#rdf.wikipathways.org!1d576fb0!groupRefrdf.wikipathways.org!4e598fd1!string>;
  void:triples "55988"^^xsd:long .

<https://rdf.wikipathway.org/.well-known/void#rdf.wikipathways.org!0460e71d!Grouprdf.wikipathways.org!ac64b4fc!hasPublicationXref>
  a void:Dataset;
  void:property <http://vocabularies.wikipathways.org/gpml#hasPublicationXref>;
  void:classPartition <https://rdf.wikipathway.org/.well-known/void#rdf.wikipathways.org!0460e71d!Grouprdf.wikipathways.org!ac64b4fc!hasPublicationXrefrdf.wikipathways.org!564d1421!PublicationXref>,
    <https://rdf.wikipathway.org/.well-known/void#rdf.wikipathways.org!0460e71d!Grouprdf.wikipathways.org!ac64b4fc!hasPublicationXrefrdf.wikipathways.org!f6bd5618!PublicationReference>;
  void:triples "608"^^xsd:long .

I kind of see the principle, but I don't see many moments where it is actually useful. Because if we want to just get the list of existing of available properties we can still just do a SELECT DISTINCT on the propertyPartitions that are linked to classPartitions, and same if we just want to get the available classes

By removing some of the void:Dataset that are calculated we might just lose some counts of properties/classes in specific contexts, but are these counts really needed?

Having so many different types of void:Dataset:

• increases the time to generate the statistics, and load on endpoints
• Increases the complexity of the generated statistics, making it scarier to maintainers
• A lot of them does not seems to add any interesting information that cannot be retrieved by a SPARQL query on the void:Dataset that are linking subjectsClass to objectsClass through predicates (if we have this we have 90% of interesting info)

Would it make sense to provide a flag to only enable calculating the following stats?

• subjectsClass to objectsClass through predicates (with distinct subjects and objects count)
• Distinct count for each class and predicates alone (that can be nice general info I guess)