$ mkvirtualenv -p $(which python3) taipan
$ cdvirtualenv
$ mkdir src && cd src
$ git clone git@github.com:dice-group/TAIPAN.git && cd TAIPAN
$ pip install -e .make dev
For subject column identification use scidentifier script, for example:
$ ./bin/scidentifier data/tables/329f559f-981f-4e7a-af8c-df8838caf74a.csv
[0]
For property recommendation use propertyrecommender script, for example:
$ ./bin/propertyrecommender data/tables/329f559f-981f-4e7a-af8c-df8838caf74a.csv
[
{
"col_i": 0,
"properties": [
{
"score": 2.039461,
"prefixed_name": "dbpedia-owl:recordLabel",
"uri": "http://dbpedia.org/ontology/recordLabel"
},
{
"score": 1.4604475,
"prefixed_name": "dbpedia-owl:distributingLabel",
"uri": "http://dbpedia.org/ontology/distributingLabel"
},
... (trimmed)
]
}
]
For SML mapping generation:
$ ./bin/smlmapping data/tables/329f559f-981f-4e7a-af8c-df8838caf74a.csv http://example.org
Prefix fn:<http://aksw.org/sparqlify/>
Prefix dbpedia-owl:<http://dbpedia.org/ontology/>
Create View Template csvtemplate As
Construct {
?subjectVariable dbpedia-owl:recordLabel ?columnVariable1 .
}
With
?subjectVariable = uri(concat("http://example.org", fn:urlEncode(?1)))
?columnVariable1 = plainLiteral(?1)
You can use mappings together with SPARQLify to convert csv to RDF:
$ ./bin/smlmapping data/tables/329f559f-981f-4e7a-af8c-df8838caf74a.csv http://example.org > generated_mapping.sml
$ sparqlify-csv -c generated_mapping.sml -f data/tables/329f559f-981f-4e7a-af8c-df8838caf74a.csv