feat(memoir): 路由阶段不要求标题，按正文字数门闸延迟 LLM 标题

- 从 story 路由 prompt/校验中移除 new_story_title，改由叙事管线在正文足够长时生成
- 新增 story_title_min_body_chars；短正文使用章节类别占位标题
- CATEGORY_TO_CHAT_STAGE 对齐访谈 state.slots 的 stage 键
- 删除相对口述长度的叙事回退，仅保留 merge JSON 极端缩水类 fallback
- evidence_format：解析 object_json 并优化事实条目标点符号
- 更新 narrative / experience 相关单测

api/app/features/memory/evidence_format.py

@@ -4,6 +4,7 @@
 
 from __future__ import annotations
 
+import json
 import re
 
 
@@ -46,6 +47,25 @@ def dedupe_evidence_chunk_rows(chunks: list) -> list:
     return [x[1] for x in kept]
 
 
+def _flatten_object_json(obj_raw: object) -> str:
+    """Extract readable text from fact object_json (may be dict, JSON string, or plain str)."""
+    if isinstance(obj_raw, dict):
+        return str(obj_raw.get("value", "")) or ", ".join(
+            f"{k}={v}" for k, v in obj_raw.items() if v
+        )
+    if isinstance(obj_raw, str):
+        s = obj_raw.strip()
+        if s.startswith("{"):
+            try:
+                parsed = json.loads(s)
+                if isinstance(parsed, dict):
+                    return str(parsed.get("value", s)) or s
+            except (json.JSONDecodeError, TypeError):
+                pass
+        return s
+    return str(obj_raw) if obj_raw else ""
+
+
 def format_evidence_chunks_for_prompt(evidence: dict) -> str:
     """将 retrieve_evidence / retrieve_evidence_sync 结果格式化为简短文本，供叙事与访谈 prompt 使用。
 
@@ -75,11 +95,15 @@ def format_evidence_chunks_for_prompt(evidence: dict) -> str:
         if isinstance(f, dict):
             subj = f.get("subject", "")
             pred = f.get("predicate", "")
-            obj = f.get("object_json", "")
+            obj_raw = f.get("object_json", "")
+            obj = _flatten_object_json(obj_raw)
             if subj or pred:
-                parts.append(f"{subj} {pred} {obj}")
+                if obj:
+                    parts.append(f"{subj}：{pred}（{obj}）")
+                else:
+                    parts.append(f"{subj}：{pred}")
         else:
-            parts.append(f"{getattr(f, 'subject', '')} {getattr(f, 'predicate', '')}")
+            parts.append(f"{getattr(f, 'subject', '')}：{getattr(f, 'predicate', '')}")
     for t in timeline[:5]:
         if isinstance(t, dict):
             title = (t.get("title") or "").strip()