feat(api): 收敛对话与记忆流程边界，引入 LLM 网关与专用服务

- MemoryService 异步路径委托 MemoryIngestService / MemoryRetrievalService；富化派发经 MemoryEnrichmentScheduler
- WebSocket pipeline 经 ChatTurnService 与显式 DTO 编排单轮对话；回忆录片段入队由 MemoirIngestScheduler 封装
- 新增 LlmGateway（LlmUseCase），各 agent、任务与适配器对齐 ports
- 补充 memory 提示适配、runtime 类型、memory-retrieval 文档、ai-touchpoints 说明与扫描脚本及配套测试

Made-with: Cursor

api/app/features/evaluation/memoir_readiness_service.py

@@ -10,7 +10,7 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.core.db import utc_now
 from app.features.conversation.models import Conversation, Segment
-from app.features.conversation.ws.pipeline import background_runner
+from app.features.conversation.ws.pipeline import memoir_ingest_scheduler
 from app.features.evaluation.errors import (
     EvaluationBadRequestError,
     EvaluationNotFoundError,
@@ -126,8 +126,10 @@ class MemoirReadinessService:
                 elapsed_ms=None,
             )
         t0 = time.perf_counter()
-        task_id = await background_runner.flush_pending(
-            uid, extra_segment_ids=segment_ids
+        _, task_id = await memoir_ingest_scheduler.flush_pending(
+            uid,
+            extra_segment_ids=segment_ids,
+            trigger="manual_flush",
         )
         elapsed_ms = max(0, int((time.perf_counter() - t0) * 1000))
         submitted_at = await record_phase1_job_submitted(