Add author_cell_type_annotations (JSON) per Cell_Cluster

[ubyndr]

Consolidate author cell-type annotations into one JSON blob per CellSet and attach it via the RDF predicate ns.author_cell_type_annotations.

Why

• Simpler, key-agnostic querying downstream (esp. Neo4j with APOC).
• Fewer triples and less schema churn than per-key predicates.
• Keeps a single namespace (self.ns).

What changed

• Build a payload from inner_dict including only author keys.

• After the per-resource loop, write one triple:

  self.graph.add(
      (resource,
       self.ns.author_cell_type_annotations,
       Literal(json.dumps(payload, ensure_ascii=False)))
  )

Before vs After

Before (multiple triples):

ex:cell a ex:CellSet ;
  ex:subclass.full "Cycling Connecting Tubule Cell" ;
  ex:subclass.l2   "cycCNT" ;
  ex:subclass.l3   "cycCNT" .

After (single JSON triple):

ex:cell a ex:CellSet ;
  ex:author_cell_type_annotations
    "{\"subclass.full\":\"Cycling Connecting Tubule Cell\",\"subclass.l2\":\"cycCNT\",\"subclass.l3\":\"cycCNT\"}" .