feat(api): 访谈人格/回复长度策略、口述归一、背景语气与输入净稿全链路

Chat 访谈
- 新增 persona 系统（default / warm_listener / curious_guide）与 background_voice 语气层
- 回复长度由 compute_reply_plan 统一决策（brief / standard / expanded），融合信息密度启发式
- 输入净稿（input_normalize）：编排层可选 rules/llm 归一用户口语后再喂模型与记忆检索
- 记忆证据注入：按用户话检索 memory evidence 并注入 prompt

Memoir 回忆录
- 口述归一（oral_normalize）：segment 原文保留，story 管线取派生净稿作叙事输入
- segment 入队批次门闸：累计字数 + 最长等待秒数，减少零碎提交
- fidelity_check / prompts / narrative_agent 微调
- Alembic 0005：清理跨章节 story 外键

Infra
- Dockerfile 加入 ffmpeg
- pyproject.toml 新增依赖并同步 uv.lock
- .env.example / .env.production 补全新配置项

Tests
- 新增 test_background_voice、test_chat_input_normalize、test_experience_regressions
- 扩展 test_interview_prompts、test_interview_reply_length、test_story_route_oral_invariant

Made-with: Cursor

api/app/agents/memoir/orchestrator.py

@@ -31,6 +31,8 @@ class PreparedMemoirBatches:
 
     state: MemoirStateSchema
     category_to_segments: Dict[str, List[Segment]]
+    #: segment id 在「LLM 判 none 且 extraction slots 为空」时加入；batch 级短路见 memoir_tasks
+    segment_skip_story_ids: Set[str]
 
 
 class MemoirOrchestrator:
@@ -58,6 +60,7 @@ class MemoirOrchestrator:
         """
         state = get_or_create_state()
         category_to_segments: Dict[str, List[Segment]] = {}
+        segment_skip_story_ids: Set[str] = set()
 
         for segment in segments:
             text = segment.user_input_text or ""
@@ -87,12 +90,16 @@ class MemoirOrchestrator:
                 "MemoirOrchestrator.ClassificationAgent.classify",
                 segment_id=segment.id,
             ):
-                chapter_category = self.classification_agent.classify(
+                classify_result = self.classification_agent.classify(
                     text=text,
                     fallback_stage=detected_stage,
                     llm=llm,
                     segment_id=segment.id,
                 )
+            chapter_category = classify_result.category
+            if (not result.slots) and classify_result.llm_said_none:
+                segment_skip_story_ids.add(str(segment.id))
+
             if agent_summary_enabled():
                 logger.info(
                     "MemoirOrchestrator.segment segment_id={} text_len={} "
@@ -114,6 +121,7 @@ class MemoirOrchestrator:
         return PreparedMemoirBatches(
             state=state,
             category_to_segments=category_to_segments,
+            segment_skip_story_ids=segment_skip_story_ids,
         )
 
     def run(