Fix: prevent VertexAiRagRetrieval from blocking event loop

Author: loic-combis

src/google/adk/tools/retrieval/vertex_ai_rag_retrieval.py

@@ -17,6 +17,7 @@
 from __future__ import annotations
 
 import logging
+import asyncio
 from typing import Any
 from typing import TYPE_CHECKING
 
@@ -94,13 +95,13 @@ async def run_async(
   ) -> Any:
     from ...dependencies.vertexai import rag
 
-    response = rag.retrieval_query(
+    response = asyncio.to_thread(rag.retrieval_query(
         text=args['query'],
         rag_resources=self.vertex_rag_store.rag_resources,
         rag_corpora=self.vertex_rag_store.rag_corpora,
         similarity_top_k=self.vertex_rag_store.similarity_top_k,
         vector_distance_threshold=self.vertex_rag_store.vector_distance_threshold,
-    )
+    ))
 
     logging.debug('RAG raw response: %s', response)