[BREAKING] Python: Move single-config fluent methods to constructor parameters (#3693)

* Move single-config fluent methods to constructor parameters

* Updates

* Adjust magentic and group chat

python/packages/lab/lightning/tests/test_lightning.py

@@ -97,10 +97,7 @@ def workflow_two_agents():
 
             # Build workflow: analyzer -> advisor
             workflow = (
-                WorkflowBuilder()
-                .set_start_executor(analyzer_executor)
-                .add_edge(analyzer_executor, advisor_executor)
-                .build()
+                WorkflowBuilder(start_executor=analyzer_executor).add_edge(analyzer_executor, advisor_executor).build()
             )
 
             yield workflow

python/packages/lab/tau2/README.md

@@ -165,15 +165,12 @@ from agent_framework.lab.tau2 import TaskRunner
 
 class WorkflowTaskRunner(TaskRunner):
     def build_conversation_workflow(self, assistant_agent, user_simulator_agent):
-        # Build a custom workflow
-        builder = WorkflowBuilder()
-
         # Create agent executors
         assistant_executor = AgentExecutor(assistant_agent, id="assistant_agent")
         user_executor = AgentExecutor(user_simulator_agent, id="user_simulator")
 
-        # Add workflow edges and conditions
-        builder.set_start_executor(assistant_executor)
+        # Build a custom workflow with start executor
+        builder = WorkflowBuilder(start_executor=assistant_executor)
         builder.add_edge(assistant_executor, user_executor)
         builder.add_edge(user_executor, assistant_executor, condition=self.should_not_stop)
 

python/packages/lab/tau2/agent_framework_lab_tau2/runner.py

@@ -288,8 +288,8 @@ class TaskRunner:
         # Creates a cyclic workflow: Orchestrator -> Assistant -> Orchestrator -> User -> Orchestrator...
         # The orchestrator acts as a message router that flips roles and routes to appropriate agent
         return (
-            WorkflowBuilder(max_iterations=10000)  # Unlimited - we control termination via should_not_stop
-            .set_start_executor(orchestrator)  # Orchestrator manages the conversation flow
+            # Orchestrator manages the conversation flow
+            WorkflowBuilder(max_iterations=10000, start_executor=orchestrator)
             .add_edge(orchestrator, self._assistant_executor)  # Route messages to assistant
             .add_edge(
                 self._assistant_executor, orchestrator, condition=self.should_not_stop