feat(memory,conversation): 记忆富化/证据包、时间线幂等字段与对话分段全链路

数据库
- 新增迁移 0003：timeline_events.memory_source_id 外键 → memory_sources，便于按 ingest 源做时间线幂等

后端 - 记忆
- 新增 ingest 后 LLM 富化（摘要/事实/时间线），可配置开关与最大字符数
- 新增证据包组装：合并 chunk、摘要、事实、时间线、故事等检索结果；支持空 query 时是否仍带 rolling 等开关
- repo/retriever/service/router/schemas/summarizer/timeline/extractor 等扩展；文档 memory-retrieval.md 更新

后端 - 对话 WS
- 增加 PING/PONG；分段 ASR 日志与空音频处理；转写失败与「无助手回复」错误提示更明确
- 助手多段回复持久化使用统一分隔符，与分段逻辑一致

后端 - Agent
- reply_limits：按 [SPLIT] 与段落拆段，并保证非空 fallback，供 WS 与 TTS 多段下发

后端 - 回忆录任务
- transcript ingest 记录 source_id；任务成功结?

api/app/features/conversation/history_store.py

@@ -18,6 +18,9 @@ from app.features.conversation.session_history import (
 
 logger = get_logger(__name__)
 
+# 与 LLM / 客户端约定：多段助手消息用 [SPLIT] 拼接，便于拆成多条气泡与多段 TTS
+AI_RESPONSE_SEGMENT_JOIN = "[SPLIT]"
+
 
 def _utc_now() -> datetime:
     return datetime.now(timezone.utc)
@@ -55,10 +58,10 @@ class ConversationHistoryStore:
 
     async def record_ai_only_turn(
         self, conversation_id: str, responses: list[str]
-    ) -> None:
+    ) -> str | None:
         if not responses:
-            return
-        combined = "\n\n".join(responses)
+            return None
+        combined = AI_RESPONSE_SEGMENT_JOIN.join(responses)
         created_at = _utc_now()
         msg = ConversationMessage(
             id=str(uuid.uuid4()),
@@ -72,6 +75,7 @@ class ConversationHistoryStore:
         await self._touch_conversation(conversation_id, occurred_at=created_at)
         await self._db.commit()
         await self._sync_redis_best_effort(conversation_id)
+        return msg.id
 
     async def record_human_ai_turn(
         self,
@@ -106,7 +110,7 @@ class ConversationHistoryStore:
             segment_id=segment_id,
             created_at=human_ts,
         )
-        combined = "\n\n".join(responses)
+        combined = AI_RESPONSE_SEGMENT_JOIN.join(responses)
         ai = ConversationMessage(
             id=str(uuid.uuid4()),
             conversation_id=conversation_id,