修复：CI 部署环境与 ref 错配、迁移碎片化、图片意图 source_span、章节物化脏版式、会话历史与本地语音不一致

新增：TTS 上传 COS 与分片、章节 reading_segments 物化与快照、markdown 清洗、会话消息 repository、语音 store 重构与相关测试

api/app/agents/chat/conversation_agent.py

@@ -39,6 +39,7 @@ class ConversationAgent:
         is_from_voice: bool = False,
         voice_session_id: str | None = None,
         user_message_timestamp: datetime | None = None,
+        audio_duration_seconds: int | None = None,
     ) -> List[str]:
         """委托 ChatOrchestrator/ProfileAgent 生成资料追问"""
         return await self._orchestrator.generate_profile_followup(
@@ -50,6 +51,7 @@ class ConversationAgent:
             is_from_voice=is_from_voice,
             voice_session_id=voice_session_id,
             user_message_timestamp=user_message_timestamp,
+            audio_duration_seconds=audio_duration_seconds,
         )
 
     async def generate_profile_greeting(
@@ -74,6 +76,7 @@ class ConversationAgent:
         is_from_voice: bool = False,
         voice_session_id: str | None = None,
         user_message_timestamp: datetime | None = None,
+        audio_duration_seconds: int | None = None,
     ) -> List[str]:
         """委托 ChatOrchestrator/InterviewAgent 生成访谈回复"""
         return await self._orchestrator.generate_response_with_state(
@@ -84,6 +87,7 @@ class ConversationAgent:
             is_from_voice=is_from_voice,
             voice_session_id=voice_session_id,
             user_message_timestamp=user_message_timestamp,
+            audio_duration_seconds=audio_duration_seconds,
         )
 
     async def generate_opening_message(