Embeddings API scores (#671)

- Put scores in all responses
- Remove unused 'middle' vector layer. Vector of texts -> vector of (vector embedding)

tests/integration/test_graph_rag_integration.py

@@ -11,6 +11,7 @@ NOTE: This is the first integration test file for GraphRAG (previously had only
 import pytest
 from unittest.mock import AsyncMock, MagicMock
 from trustgraph.retrieval.graph_rag.graph_rag import GraphRag
+from trustgraph.schema import EntityMatch, Term, IRI
 
 
 @pytest.mark.integration
@@ -35,9 +36,9 @@ class TestGraphRagIntegration:
         """Mock graph embeddings client that returns realistic entities"""
         client = AsyncMock()
         client.query.return_value = [
-            "http://trustgraph.ai/e/machine-learning",
-            "http://trustgraph.ai/e/artificial-intelligence",
-            "http://trustgraph.ai/e/neural-networks"
+            EntityMatch(entity=Term(type=IRI, iri="http://trustgraph.ai/e/machine-learning"), score=0.95),
+            EntityMatch(entity=Term(type=IRI, iri="http://trustgraph.ai/e/artificial-intelligence"), score=0.90),
+            EntityMatch(entity=Term(type=IRI, iri="http://trustgraph.ai/e/neural-networks"), score=0.85)
         ]
         return client
 
@@ -130,7 +131,7 @@ class TestGraphRagIntegration:
         # 2. Should query graph embeddings to find relevant entities
         mock_graph_embeddings_client.query.assert_called_once()
         call_args = mock_graph_embeddings_client.query.call_args
-        assert call_args.kwargs['vectors'] == [[0.1, 0.2, 0.3, 0.4, 0.5]]
+        assert call_args.kwargs['vector'] == [[0.1, 0.2, 0.3, 0.4, 0.5]]
         assert call_args.kwargs['limit'] == entity_limit
         assert call_args.kwargs['user'] == user
         assert call_args.kwargs['collection'] == collection