- 新增 utterance_substance:短时/应答/元话语可跳过记忆检索、阶段 LLM 与资料抽取 LLM;可配置 - 输入归一化:LLM 模式默认仅语音/ASR;配置项写入 .env.example - Memoir Phase1:可选 batch LLM 一次性抽取+分类(失败回退逐段);Extraction 空槽位时阶段与 current_stage 对齐,prompt 约束收紧 - 叙事与忠实度:narrative_safety、证据重叠/场合锚点、标题 slots 与履历短语 grounded;fidelity 解析失败 fail-open 可配置 - 章节管线:锁 TTL 上调、锁竞争 Celery 重试、Phase2 immediate singleflight 等;story_pipeline_sync / chapter_compose / memoir_tasks 联动 - Memory:compaction / repo / summarizer / evidence 小修;事实 FTS 未命中是否回退最近事实可配置 - 新增 memoir_pipeline_trace;补充 memoir_reliability 文档与多项回归/门控测试
35 lines
1.2 KiB
Python
35 lines
1.2 KiB
Python
"""将 NarrativeAgent / LLM 返回的 JSON 或纯文本规范为 markdown 正文。"""
|
||
|
||
from __future__ import annotations
|
||
|
||
import json
|
||
|
||
|
||
def narrative_to_markdown(narrative: str) -> str:
|
||
"""
|
||
将 narrative(JSON paragraphs 或纯文本)转为 markdown。
|
||
与已删除的 ChapterComposerOrchestrator._to_markdown 行为一致。
|
||
"""
|
||
if not narrative or not str(narrative).strip():
|
||
return ""
|
||
stripped = narrative.strip()
|
||
if stripped.startswith("{") and "paragraphs" in stripped:
|
||
try:
|
||
data = json.loads(stripped)
|
||
paras = data.get("paragraphs", [])
|
||
if isinstance(paras, list):
|
||
parts = []
|
||
for p in paras:
|
||
if isinstance(p, dict):
|
||
text = p.get("content", p.get("text", ""))
|
||
else:
|
||
text = str(p)
|
||
if text.strip():
|
||
parts.append(text.strip())
|
||
return "\n\n".join(parts)
|
||
return stripped
|
||
except json.JSONDecodeError:
|
||
# 不得以伪 JSON 字符串落库;上层 _coalesce_story_markdown 会回退口述/旧文
|
||
return ""
|
||
return stripped
|