refactor(api,expo): 多智能体与会话收敛、回忆录兼容层移除、后端测试集大幅删减

- 对齐「多智能体收敛」与「回忆录 stories-first / markdown-first」方向：收紧运行时契约、
  删除过渡兼容路径与双轨逻辑，并同步更新客户端与文档。

- Chat：以 ChatOrchestrator 为实时编排入口；删除独立 conversation_agent，精简 prompts。
- Memoir：删除 memory_agent；MemoirOrchestrator、classification / story_route 与 prompts 收敛到
  prepare_batches + run_story_pipeline_for_category_batch 主链路。
- 将 agents 侧 processor 迁入 feature 层为 background_runner，并移除 features 下重复/过时
  processor 封装。

- 新增 history_store，强化「conversation_messages 为 DB 真源、Redis 为缓存」模型。
- 调整 models、repo、service、session_history；精简 WS message_types，重构 pipeline 与 router。

- 移除章节占位、整章再生等旧路径；章节列表与封面逻辑要求 story 关联；收紧 cover 资格与
  enqueue。
- helpers、repo、service、router、reading_segment_materialize、story_pipeline_sync、pdf_service
  等按 canonical markdown / cover_asset_id 收缩；删除 memoir_images/provider 等冗余。
- tasks：memoir_tasks、chapter_cover_tasks 等大幅瘦身；story_image_tasks 等与当前图片任务对齐。

- core：config、logging、redis、task_tracker 小幅调整。
- auth / user / payment / quota：路由或服务侧删减过时接口或逻辑（如 payment router 行数减少）。

- pyproject.toml、development.sh、.env.example / .env.production、README 等同步说明或变量。

- Alembic 0001_initial_schema 微调（与当前 schema 叙事一致的小改动）。

- 回忆录：types / mappers / api、章节页与 memoir 页与后端契约对齐；markdown-renderer 调整。
- 语音：删除 voice/player，voice-segment-store 相应精简。

- api/tests：删除 conftest 及绝大部分既有测试文件（websocket_baseline、conversation、memoir
  图片、PDF、SMS 等），属有意收缩/待按 backend-test-system 重建的信号。
- docs：新增多智能体收敛与移除兼容层计划摘要；更新 story-first 设计、backend-test-system、
  multi-agent-refactor-plan、实施总结等。

BREAKING CHANGE: 后端对外契约、回忆录章节字段与若干路由/任务行为已变更；大量 API 测试被移除，
  CI 若依赖这些用例需按新策略补测或调整流水线。

api/app/agents/chat/__init__.py

@@ -1,12 +1,10 @@
 """聊天模块：AI 回复用户（ProfileAgent + InterviewAgent + ChatOrchestrator）"""
 
-from app.agents.chat.conversation_agent import ConversationAgent
+from app.agents.chat.interview_agent import InterviewAgent
 from app.agents.chat.orchestrator import ChatOrchestrator
 from app.agents.chat.profile_agent import ProfileAgent
-from app.agents.chat.interview_agent import InterviewAgent
 
 __all__ = [
-    "ConversationAgent",
     "ChatOrchestrator",
     "ProfileAgent",
     "InterviewAgent",

api/app/agents/chat/conversation_agent.py

@@ -1,147 +0,0 @@
-"""
-对话 Agent：Facade，内部委托 ChatOrchestrator + ProfileAgent + InterviewAgent
-保留原有对外 API，供 router 等调用方兼容使用
-"""
-
-from datetime import datetime
-from typing import Any, Dict, List, Optional
-
-from app.agents.chat.agent_turn import AgentChatTurn
-from app.agents.chat.orchestrator import ChatOrchestrator
-from app.agents.chat.prompts_conversation import ConversationStage
-from app.agents.state_schema import MemoirStateSchema
-from app.core.redis import redis_service
-
-
-class ConversationAgent:
-    """对话 Agent Facade，委托 ChatOrchestrator 实现多 Agent 协同"""
-
-    def __init__(self):
-        self._orchestrator = ChatOrchestrator()
-
-    async def extract_profile_from_message(
-        self,
-        user_message: str,
-        missing_fields: List[str],
-        conversation_id: Optional[str] = None,
-    ) -> Dict[str, Any]:
-        """委托 ChatOrchestrator/ProfileAgent 提取资料"""
-        return await self._orchestrator.extract_profile_from_message(
-            user_message, missing_fields, conversation_id=conversation_id
-        )
-
-    async def generate_profile_followup(
-        self,
-        conversation_id: str,
-        user_message: str,
-        missing_fields: List[str],
-        filled_fields: Dict[str, str],
-        nickname: str = "",
-        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(
-            conversation_id=conversation_id,
-            user_message=user_message,
-            missing_fields=missing_fields,
-            filled_fields=filled_fields,
-            nickname=nickname,
-            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(
-        self,
-        conversation_id: str,
-        missing_fields: List[str],
-        nickname: str = "",
-    ) -> List[str]:
-        """委托 ChatOrchestrator/ProfileAgent 生成资料收集开场白"""
-        return await self._orchestrator.generate_profile_greeting(
-            conversation_id=conversation_id,
-            missing_fields=missing_fields,
-            nickname=nickname,
-        )
-
-    async def generate_response_with_state(
-        self,
-        conversation_id: str,
-        user_message: str,
-        memoir_state: MemoirStateSchema,
-        user_profile_context: str = "",
-        is_from_voice: bool = False,
-        voice_session_id: str | None = None,
-        user_message_timestamp: datetime | None = None,
-        audio_duration_seconds: int | None = None,
-    ) -> AgentChatTurn:
-        """委托 ChatOrchestrator/InterviewAgent 生成访谈回复"""
-        return await self._orchestrator.generate_response_with_state(
-            conversation_id=conversation_id,
-            user_message=user_message,
-            memoir_state=memoir_state,
-            user_profile_context=user_profile_context,
-            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(
-        self,
-        conversation_id: str,
-        memoir_state: MemoirStateSchema,
-        user_profile_context: str = "",
-    ) -> List[str]:
-        """委托 ChatOrchestrator/InterviewAgent 生成开场白"""
-        return await self._orchestrator.generate_opening_message(
-            conversation_id=conversation_id,
-            memoir_state=memoir_state,
-            user_profile_context=user_profile_context,
-        )
-
-    async def generate_response(
-        self,
-        conversation_id: str,
-        user_message: str,
-        current_stage: Optional[ConversationStage] = None,
-        covered_topics: Optional[List[str]] = None,
-    ) -> str:
-        """兼容旧 API：生成简单回复（无状态感知），委托 InterviewAgent 的等价逻辑"""
-        from app.agents.state_schema import default_state
-
-        state = default_state()
-        state.current_stage = (current_stage or ConversationStage.CHILDHOOD).value
-        state.covered_stages = covered_topics or []
-        turn = await self._orchestrator.generate_response_with_state(
-            conversation_id=conversation_id,
-            user_message=user_message,
-            memoir_state=state,
-            user_profile_context="",
-        )
-        return turn.messages[0] if turn.messages else ""
-
-    def detect_stage(
-        self, conversation_id: str, user_message: str
-    ) -> ConversationStage:
-        """根据关键词检测用户阶段（兼容 API）"""
-        detected = self._orchestrator.detect_user_stage(user_message)
-        if detected == "childhood":
-            return ConversationStage.CHILDHOOD
-        if detected == "education":
-            return ConversationStage.EDUCATION
-        if detected == "career":
-            return ConversationStage.CAREER
-        if detected == "family":
-            return ConversationStage.FAMILY
-        if detected == "belief":
-            return ConversationStage.BELIEFS
-        return ConversationStage.CHILDHOOD
-
-    async def clear_memory(self, conversation_id: str) -> None:
-        """清除 Redis 中的对话历史"""
-        await redis_service.clear_conversation_history(conversation_id)

api/app/agents/chat/orchestrator.py

@@ -1,6 +1,6 @@
 """
 ChatOrchestrator：AI 回复用户模块的编排层
-负责路由（Profile vs Interview）、调用 Specialist Agent、统一 Redis 持久化与错误处理
+负责路由（Profile vs Interview）、调用 Specialist Agent；持久化由 feature 层 ConversationHistoryStore 完成。
 """
 
 from datetime import datetime
@@ -9,7 +9,6 @@ from typing import TYPE_CHECKING, List, Optional
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.agents.chat.agent_turn import AgentChatTurn
-from app.agents.chat.helpers import save_message
 from app.agents.chat.interview_agent import InterviewAgent
 from app.agents.chat.profile_agent import ProfileAgent
 from app.agents.state_schema import MemoirStateSchema
@@ -28,8 +27,8 @@ _UNAUTH_TURN = AgentChatTurn(
 
 class ChatOrchestrator:
     """
-    聊天编排器：根据用户资料完成度路由到 ProfileAgent 或 InterviewAgent，
-    统一管理 Redis 写入。
+    聊天编排器：根据用户资料完成度路由到 ProfileAgent 或 InterviewAgent。
+    不直接写入 Redis/DB；由 WS pipeline / ConversationHistoryStore 落库并同步缓存。
     """
 
     def __init__(self):
@@ -53,8 +52,7 @@ class ChatOrchestrator:
     ) -> AgentChatTurn:
         """
         处理用户消息，返回 AI 回复（分段 + 是否跳过 TTS）。
-        根据 missing_fields 路由到 ProfileAgent 或 InterviewAgent，
-        统一写入 Redis。
+        根据 missing_fields 路由到 ProfileAgent 或 InterviewAgent。
         """
 
         # --- 资料收集模式 ---
@@ -77,15 +75,6 @@ class ChatOrchestrator:
                         filled_fields=filled,
                         nickname=user.nickname or "",
                     )
-                    await self._save_messages(
-                        conversation_id=conversation_id,
-                        user_message=user_message,
-                        response_text="\n\n".join(responses),
-                        is_from_voice=is_from_voice,
-                        voice_session_id=voice_session_id,
-                        user_message_timestamp=user_message_timestamp,
-                        audio_duration_seconds=audio_duration_seconds,
-                    )
                     return AgentChatTurn(messages=responses, skip_tts=False)
                 except Exception as e:
                     logger.error(f"资料收集处理失败: {e}", exc_info=True)
@@ -111,52 +100,12 @@ class ChatOrchestrator:
                 occupation=user.occupation,
             )
 
-        turn = await self.interview_agent.generate_response_with_state(
+        return await self.interview_agent.generate_response_with_state(
             conversation_id=conversation_id,
             user_message=user_message,
             memoir_state=state,
             user_profile_context=user_profile_context,
         )
-        await self._save_messages(
-            conversation_id=conversation_id,
-            user_message=user_message,
-            response_text="\n\n".join(turn.messages),
-            is_from_voice=is_from_voice,
-            voice_session_id=voice_session_id,
-            user_message_timestamp=user_message_timestamp,
-            audio_duration_seconds=audio_duration_seconds,
-        )
-        return turn
-
-    async def _save_messages(
-        self,
-        conversation_id: str,
-        user_message: str,
-        response_text: str,
-        is_from_voice: bool = False,
-        voice_session_id: Optional[str] = None,
-        user_message_timestamp: Optional[datetime] = None,
-        audio_duration_seconds: Optional[int] = None,
-    ) -> None:
-        """统一写入 Human + AI 消息到 Redis"""
-        human_msg_type = "audio" if is_from_voice else "text"
-        human_duration = (
-            audio_duration_seconds
-            if is_from_voice
-            and audio_duration_seconds is not None
-            and audio_duration_seconds > 0
-            else None
-        )
-        await save_message(
-            conversation_id,
-            "human",
-            user_message,
-            message_type=human_msg_type,
-            voice_session_id=voice_session_id,
-            timestamp=user_message_timestamp,
-            audio_duration_seconds=human_duration,
-        )
-        await save_message(conversation_id, "ai", response_text)
 
     async def extract_profile_from_message(
         self,
@@ -181,25 +130,14 @@ class ChatOrchestrator:
         user_message_timestamp: datetime | None = None,
         audio_duration_seconds: int | None = None,
     ) -> List[str]:
-        """委托 ProfileAgent 生成资料追问，并写入 Redis"""
-        responses = await self.profile_agent.generate_profile_followup(
+        """委托 ProfileAgent 生成资料追问（持久化由调用方负责）。"""
+        return await self.profile_agent.generate_profile_followup(
             conversation_id=conversation_id,
             user_message=user_message,
             missing_fields=missing_fields,
             filled_fields=filled_fields,
             nickname=nickname,
         )
-        response_text = "\n\n".join(responses)
-        await self._save_messages(
-            conversation_id=conversation_id,
-            user_message=user_message,
-            response_text=response_text,
-            is_from_voice=is_from_voice,
-            voice_session_id=voice_session_id,
-            user_message_timestamp=user_message_timestamp,
-            audio_duration_seconds=audio_duration_seconds,
-        )
-        return responses
 
     async def generate_profile_greeting(
         self,
@@ -207,15 +145,12 @@ class ChatOrchestrator:
         missing_fields: List[str],
         nickname: str = "",
     ) -> List[str]:
-        """委托 ProfileAgent 生成资料收集开场白，并写入 Redis"""
-        responses = await self.profile_agent.generate_profile_greeting(
+        """委托 ProfileAgent 生成资料收集开场白（持久化由调用方负责）。"""
+        return await self.profile_agent.generate_profile_greeting(
             conversation_id=conversation_id,
             missing_fields=missing_fields,
             nickname=nickname,
         )
-        response_text = "\n\n".join(responses)
-        await save_message(conversation_id, "ai", response_text)
-        return responses
 
     async def generate_response_with_state(
         self,
@@ -228,24 +163,13 @@ class ChatOrchestrator:
         user_message_timestamp: datetime | None = None,
         audio_duration_seconds: int | None = None,
     ) -> AgentChatTurn:
-        """委托 InterviewAgent 生成访谈回复，并写入 Redis"""
-        turn = await self.interview_agent.generate_response_with_state(
+        """委托 InterviewAgent 生成访谈回复（持久化由调用方负责）。"""
+        return await self.interview_agent.generate_response_with_state(
             conversation_id=conversation_id,
             user_message=user_message,
             memoir_state=memoir_state,
             user_profile_context=user_profile_context,
         )
-        response_text = "\n\n".join(turn.messages)
-        await self._save_messages(
-            conversation_id=conversation_id,
-            user_message=user_message,
-            response_text=response_text,
-            is_from_voice=is_from_voice,
-            voice_session_id=voice_session_id,
-            user_message_timestamp=user_message_timestamp,
-            audio_duration_seconds=audio_duration_seconds,
-        )
-        return turn
 
     def detect_user_stage(self, user_message: str) -> str:
         """委托 InterviewAgent 检测用户阶段"""
@@ -258,16 +182,10 @@ class ChatOrchestrator:
         user_profile_context: str = "",
     ) -> List[str]:
         """
-        委托 InterviewAgent 生成访谈开场白，并写入 Redis。
-
-        调用方（如 WS）须在「空会话」分支前通过 ConversationService 从 DB 回填 Redis，
-        避免与多 Agent 契约混淆：本编排器不读取 segments，只假定 Redis 已反映是否已有轮次。
+        委托 InterviewAgent 生成访谈开场白（持久化由调用方 ConversationHistoryStore 负责）。
         """
-        responses = await self.interview_agent.generate_opening_message(
+        return await self.interview_agent.generate_opening_message(
             conversation_id=conversation_id,
             memoir_state=memoir_state,
             user_profile_context=user_profile_context,
         )
-        response_text = "\n\n".join(responses)
-        await save_message(conversation_id, "ai", response_text)
-        return responses

api/app/agents/chat/prompts.py

@@ -17,7 +17,6 @@ from app.agents.chat.prompts_conversation import (
     ConversationStage,
     INTERVIEW_QUESTIONS,
     SLOT_NAME_MAP,
-    get_conversation_prompt,
     get_guided_conversation_prompt,
     get_opening_prompt,
     get_questions_for_stage,
@@ -34,7 +33,6 @@ __all__ = [
     "ConversationStage",
     "INTERVIEW_QUESTIONS",
     "SLOT_NAME_MAP",
-    "get_conversation_prompt",
     "get_guided_conversation_prompt",
     "get_opening_prompt",
     "get_questions_for_stage",

api/app/agents/chat/prompts_conversation.py

@@ -465,12 +465,3 @@ def get_guided_conversation_prompt(
 直接输出你要说的话（多条消息用 [SPLIT] 分隔）："""
 
     return prompt
-
-
-def get_conversation_prompt(
-    current_stage: ConversationStage,
-    covered_topics: List[str],
-    user_latest_response: str,
-) -> str:
-    """向后兼容的函数"""
-    return get_system_prompt(current_stage, covered_topics, user_latest_response)