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/ws/pipeline.py

@@ -22,7 +22,10 @@ from app.core.config import settings
 from app.core.cos_url_keys import TTS_PRESIGNED_EXPIRES_SEC
 from app.core.db import AsyncSessionLocal
 from app.core.dependencies import get_asr_provider, get_object_storage, get_tts_provider
-from app.features.conversation.history_store import ConversationHistoryStore
+from app.features.conversation.history_store import (
+    AI_RESPONSE_SEGMENT_JOIN,
+    ConversationHistoryStore,
+)
 from app.features.conversation.models import Conversation, Segment
 from app.features.conversation.ws.connection_manager import manager
 from app.features.conversation.ws.message_types import MessageType
@@ -369,6 +372,16 @@ async def process_audio_segment(
 ) -> None:
     """分段语音的异步处理：并行 ASR + 幂等落库 + 有序聚合触发 Agent。"""
     state = get_or_create_segment_state(conversation_id, voice_session_id)
+    logger.info(
+        "process_audio_segment 开始: conversation_id={} voice_session_id={} "
+        "segment_index={} is_last={} duration_s={} audio_b64_len={}",
+        conversation_id,
+        voice_session_id,
+        segment_index,
+        is_last,
+        audio_duration,
+        len(audio_base64 or ""),
+    )
 
     try:
         async with AsyncSessionLocal() as db:
@@ -420,6 +433,12 @@ async def process_audio_segment(
                 audio_bytes = base64.b64decode(audio_base64)
             except Exception:
                 audio_bytes = b""
+            if not audio_bytes:
+                logger.warning(
+                    "process_audio_segment: 解码后音频为空 conversation_id={} segment_index={}",
+                    conversation_id,
+                    segment_index,
+                )
             transcript_text = await get_asr_provider().transcribe(
                 audio_bytes, format="m4a"
             )
@@ -440,12 +459,19 @@ async def process_audio_segment(
             )
 
             if _is_transcribe_failure(transcript_text):
+                detail = (transcript_text or "").strip()
+                if detail.startswith("转写失败"):
+                    user_msg = f"分段 {segment_index} {detail}"
+                elif not detail:
+                    user_msg = f"分段 {segment_index} 转写失败：未识别到内容（请检查后端 ASR 配置）"
+                else:
+                    user_msg = f"分段 {segment_index} 转写失败：{detail[:400]}"
                 await manager.send_message(
                     conversation_id,
                     {
                         "type": MessageType.ERROR,
                         "data": {
-                            "message": f"分段 {segment_index} 转写失败，请重试该片段",
+                            "message": user_msg,
                             "segment_index": segment_index,
                         },
                         "timestamp": datetime.now(timezone.utc).isoformat(),
@@ -553,6 +579,12 @@ async def process_user_message(
     store = ConversationHistoryStore(db)
     tts_urls: list[str] = []
     try:
+        logger.info(
+            "process_user_message 开始: conversation_id={} segment_id={} user_chars={}",
+            conversation_id,
+            segment.id,
+            len(user_message or ""),
+        )
         is_from_voice = bool(segment.audio_url)
         voice_session_id = _voice_session_id_from_audio_url(segment.audio_url)
         audio_dur = getattr(segment, "audio_duration_seconds", None)
@@ -586,7 +618,7 @@ async def process_user_message(
         responses = turn.messages
         skip_tts = turn.skip_tts
 
-        segment.agent_response = "\n\n".join(responses)
+        segment.agent_response = AI_RESPONSE_SEGMENT_JOIN.join(responses)
         _mark_conversation_active(conversation)
         ai_msg_id = await store.record_human_ai_turn(
             conversation_id=conversation_id,
@@ -600,6 +632,22 @@ async def process_user_message(
             segment_id=segment.id,
         )
         if not ai_msg_id:
+            logger.warning(
+                "process_user_message: 无有效助手段落（responses 为空），conversation_id={} segment_id={}",
+                conversation_id,
+                segment.id,
+            )
+            if conversation_id in manager.active_connections:
+                await manager.send_message(
+                    conversation_id,
+                    {
+                        "type": MessageType.ERROR,
+                        "data": {
+                            "message": "未生成回复，请重试或稍后再试",
+                        },
+                        "timestamp": datetime.now(timezone.utc).isoformat(),
+                    },
+                )
             return
 
         tts_epoch_start = _tts_epoch_value(conversation_id)
@@ -614,6 +662,7 @@ async def process_user_message(
                         "text": response_text,
                         "index": i,
                         "total": n,
+                        "assistant_message_id": ai_msg_id,
                     },
                     "timestamp": datetime.now(timezone.utc).isoformat(),
                 },