Create function for SPARQL refinement in columnRefinement transformations

Currently in LDWizard, it is not yet supported to use SPARQL queries to conduct refinement transformations. This issue proposes a feature to support retrieving values from a given input query and endpoint. The query could have the CSV row/column value injected, if the correct argument is present in the query.

For example:

const sparqlQuery = `
prefix dct: <http://purl.org/dc/terms/>
select ?refinement where
 {
   ?refinement dct:title ?title; a <https://triplydb.com/Triply/linkedmdb/vocab/Film> .
   filter(lcase(?title) = lcase(?search))
 }
`
const columnrefinement: ColumnRefinement = {
  bulk: true,
  transformation: sparqlTransform(sparqlQuery, 'https://api.nightly.triplydb.com/datasets/Triply/linkedmdb/sparql')
}

A description of the sparqlTransform implementation:

type Options = {
 refinementBinding?: string ;
 queryVariable?: string ;
}

function sparqlTransform(query: string, endpoint: string, options: Options = {}): (value: string) => Map<number, string> {
  const optionsWithDefault: Options = {...{refinementBinding: 'refinement', queryVariable: 'query'}, ...options}
  // 1. parse query and make it into a union of `batchSize` queries and bind the `?search` variable to the value of the column in the source data 
  // 2. perform fetch request on endpoint
  // 3. inject variable `options.queryVariable` into the `union` query 
  // 4. get binding `options.refinementBinding`
  // 5. match the binding to the correct row
  // 5. return the Map with the row number and string values of the bindings 
}

pldn / LDWizard

Create function for SPARQL refinement in columnRefinement transformations #152