Refactor workflow agent and routing components for improved functionality

- Introduce WorkflowAgentStage to manage agent stage configurations and enhance interaction with the workflow engine.
- Implement WorkflowEdgeEvaluator for priority-aware edge evaluation, improving routing decisions based on conditions and user turns.
- Update WorkflowBrain to handle user turns and routing more effectively, ensuring agents cannot have only one default path.
- Enhance CallEndCoordinator to track speech events and manage call termination based on queued speech.
- Add new models and output handling for workflow interactions, improving clarity and maintainability.
- Update tests to validate the new routing logic and agent behavior under various scenarios.

backend/services/pipecat/call_lifecycle.py

@@ -18,6 +18,8 @@ class CallEndCoordinator:
         self._armed = False
         self._speaking = False
         self._response_speech_started = False
+        self._tracked_speeches = 0
+        self._finish_after_tracked_speech = False
         self._finished = False
         self._reason = "completed"
 
@@ -37,6 +39,16 @@ class CallEndCoordinator:
         """Wait for the next observed bot speech to finish."""
         self._armed = True
 
+    def track_speech(self) -> None:
+        """Register one fixed utterance before its TTSSpeakFrame is queued."""
+        self._tracked_speeches += 1
+
+    async def arm_after_tracked_speech(self) -> None:
+        """Finish after every already queued fixed utterance has played."""
+        self._finish_after_tracked_speech = True
+        if self._tracked_speeches == 0:
+            await self.finish()
+
     async def finish_after_current_speech(self, *, has_text: bool) -> None:
         """Finish now if speech is absent/done, otherwise wait for its stop."""
         if not has_text:
@@ -59,7 +71,15 @@ class CallEndCoordinator:
             self._response_speech_started = True
         elif isinstance(frame, BotStoppedSpeakingFrame) and self._speaking:
             self._speaking = False
-            if self._armed:
+            if self._tracked_speeches > 0:
+                self._tracked_speeches -= 1
+            if (
+                self._finish_after_tracked_speech
+                and self._tracked_speeches == 0
+            ):
+                logger.info("所有工作流结束语播报完毕,挂断通话")
+                await self.finish()
+            elif self._armed:
                 logger.info("结束语播报完毕,挂断通话")
                 await self.finish()