Update ADK doc according to issue #1227 - 7 - Vertex AI search Dynamic Configuration

docs/tools/google-cloud/vertex-ai-search.md

@@ -26,3 +26,36 @@ of the tool, see
 ```py
 --8<-- "examples/python/snippets/tools/built-in-tools/vertexai_search.py"
 ```
+
+## Dynamic Configuration (e.g., Per-User Filtering)
+
+You can dynamically configure the Vertex AI Search tool at runtime by subclassing `VertexAiSearchTool` and overriding the `_build_vertex_ai_search_config(self, context)` method. This is useful for applying context-aware filters, such as filtering results based on the current user's ID.
+
+The `context` object provides access to the agent's state, allowing you to create customized search configurations for each request.
+
+### Example: Per-User Filtering
+
+Here is an example of how to implement a dynamic filter based on a `user_id` stored in the agent's state:
+
+```python
+from google.genai import types
+from google.adk.agents.readonly_context import ReadonlyContext
+from google.adk.tools import VertexAiSearchTool
+
+class DynamicFilterSearchTool(VertexAiSearchTool):
+  def _build_vertex_ai_search_config(
+      self, ctx: ReadonlyContext
+  ) -> types.VertexAISearch:
+    user_id = ctx.state.get('user_id')
+    return types.VertexAISearch(
+        datastore=self.data_store_id,
+        filter=f"user_id = '{user_id}'"
+    )
+```
+
+In this example:
+
+1.  `DynamicFilterSearchTool` inherits from `VertexAiSearchTool`.
+2.  The `_build_vertex_ai_search_config` method is overridden to access the agent's context.
+3.  The `user_id` is retrieved from `ctx.state`.
+4.  A `VertexAISearch` object is returned with a dynamic filter that restricts the search to documents where the `user_id` field matches the user's ID.