rdfs:label for IRIs generated by docs.table-headers

[sgoetz-brox]

Examples from https://sparql-anything.readthedocs.io/en/stable/formats/Word_Processing_Document/

docs.table-headers=false:

  rdf:_7    [ rdf:_1  [ rdf:_1  "A" ;
                        rdf:_2  "B" ;
                        rdf:_3  "C"
                      ] ;
              rdf:_2  [ rdf:_1  "A1" ;
                        rdf:_2  "B1" ;
                        rdf:_3  "C1"
                      ] ;
              rdf:_3  [ rdf:_1  "A2" ;
                        rdf:_2  "B2" ;
                        rdf:_3  "C2"
                      ]
            ]

docs.table-headers=true:

  rdf:_7    [ rdf:_1  [ xyz:A   "A1" ;
                        xyz:B   "B1" ;
                        xyz:C   "C1"
                      ] ;
              rdf:_2  [ xyz:A   "A2" ;
                        xyz:B   "B2" ;
                        xyz:C   "C2"
                      ]
            ]

When set to true, the generated RDF doesn’t contain "A", "B", and "C" as literals anymore.

I think it would be useful if something like this would be generated:

xyz:A rdfs:label "A" .
xyz:B rdfs:label "B" .
xyz:C rdfs:label "C" .

(Of course one could convert the IRI to a string and extract the label from it, but this becomes ugly if percent-encoding is involved.)

Is this maybe possible already? If not, would this be something that could be added to SPARQL Anything, or would this be in conflict with its design goals?

[luigi-asprino]

Hi,

this is something that applies to every format (the label can be generated once an xyz predicate is created). An idea could be adding a new general option (e.g. generate-predicate-labels)

[luigi-asprino]

439b7fc introduces this feature