fix: Add warm_up() calls to tests after ChatGenerator tool initialization changes (#9965)

* In tests call agent.warm_up() where needed

* Add reno

* Update wording in reno note

* Update reno note

Author: vblagoje

releasenotes/notes/chat-generator-warm-up-required-2226a973dbd56273.yaml

@@ -0,0 +1,26 @@
+---
+upgrade:
+  - |
+    Agent and LLMMessagesRouter components now require explicit warm_up() calls before run() when used in standalone
+    mode (outside of a Pipeline) with tools. This change was introduced by PR #9942 which added warm_up() methods to
+    ChatGenerator components for tool initialization.
+
+    **Who is affected:** Users running Agent or LLMMessagesRouter standalone with ChatGenerators that have tools.
+
+    **Action required:** Call warm_up() before run() or run_async():
+
+    .. code-block:: python
+
+        # Agent with tools
+        agent = Agent(llm=OpenAIChatGenerator(model="gpt-4"), tools=[my_tool])
+        agent.warm_up()  # Required in standalone mode
+        result = agent.run(messages=[ChatMessage.from_user("Hello")])
+
+        # LLMMessagesRouter with tools
+        router = LLMMessagesRouter(generator=OpenAIChatGenerator(model="gpt-4"), tools=[my_tool])
+        router.warm_up()  # Required in standalone mode
+        result = router.run(messages=[ChatMessage.from_user("Hello")])
+
+    **Note:** This requirement only applies to standalone usage. When these components are used within a Pipeline,
+    warm_up() is called automatically. We are reconsidering this design choice - see issue #9966 for discussion
+    about automatically calling warm_up() at runtime instead of requiring explicit calls.

test/components/agents/test_agent_breakpoints.py

@@ -534,6 +534,7 @@ def test_live_resume_from_tool_invoker(self, tmp_path, weather_tool):
         )
 
         try:
+            agent.warm_up()
             agent.run(messages=[ChatMessage.from_user("What's the weather in Berlin?")], break_point=agent_breakpoint)
         except BreakpointException:
             pass

test/components/routers/test_llm_messages_router.py

@@ -206,6 +206,7 @@ def test_live_run(self):
         )
 
         messages = [ChatMessage.from_user("Hello")]
+        router.warm_up()
         result = router.run(messages)
         print(result)
 

test/tools/test_pipeline_tool.py

@@ -297,6 +297,7 @@ def test_live_pipeline_tool(self):
         )
 
         # Let the Agent handle a query
+        agent.warm_up()
         result = agent.run([ChatMessage.from_user("Who was Nikola Tesla?")])
 
         assert len(result["messages"]) == 5  # System msg, User msg, Agent msg, Tool call result, Agent mgs
@@ -343,6 +344,7 @@ async def test_live_async_pipeline_tool(self):
         )
 
         # Let the Agent handle a query
+        agent.warm_up()
         result = await agent.run_async([ChatMessage.from_user("Who was Nikola Tesla?")])
 
         assert len(result["messages"]) == 5  # System msg, User msg, Agent msg, Tool call result, Agent mgs