Improve parsing for dataset #18

If concepts have multiple topConcepts of inScheme entries, it was kind of arbitrary until now, what would be chosen as the dataset.
This is now improved by setting the dataset to the concept scheme id.

src/publishToReconciliation/buildJSON.js

@@ -13,7 +13,8 @@ export const buildJSON = async (ttlString, account) => {
   const compacted = await jsonld.compact(expanded, context);
   // TODO get all available languages and store them as attribute for Concept Scheme
   let entries = [];
-  let dataset = "";
+  const dataset = compacted["@graph"].find(g => g.type === "ConceptScheme").id;
+  console.log("found dataset", dataset)
 
   compacted["@graph"].forEach((graph, _) => {
     const { ...properties } = graph;
@@ -68,7 +69,6 @@ export const buildJSON = async (ttlString, account) => {
     };
 
     if (node.type === "ConceptScheme") {
-      dataset = node.id;
       if (node?.preferredNamespaceUri === undefined) {
         throw new NoPrefNamespaceUriError(
           `ConceptScheme ${node.id} does not have a preferredNamespaceUri`
@@ -82,17 +82,8 @@ export const buildJSON = async (ttlString, account) => {
       if (node.topConceptOf) {
         node.inScheme = node.topConceptOf;
       }
-      try {
-
-        dataset = node?.inScheme?.[0]?.id ?? node.topConceptOf[0].id;
-      }
-      catch (e) {
-        console.error(e)
-        console.error(node)
-      }
     }
 
-
     node["dataset"] = dataset;
     node["account"] = account;