修复:CI 部署环境与 ref 错配、迁移碎片化、图片意图 source_span、章节物化脏版式、会话历史与本地语音不一致

新增:TTS 上传 COS 与分片、章节 reading_segments 物化与快照、markdown 清洗、会话消息 repository、语音 store 重构与相关测试
This commit is contained in:
Kevin
2026-03-20 16:36:42 +08:00
parent 7317bf10cd
commit 8af37e5e8e
65 changed files with 1704 additions and 504 deletions

View File

@@ -1,29 +1,16 @@
"""
按 chapter_story_links 顺序将各 story 正文物化为单一 markdown无 LLM
保留 story 内 asset:// 引用不变。
章节级 canonical仅正文拼接故事间用 ---;故事标题仅存 stories.title。
PDF 导出可单独物化「## 标题 + 正文」版本。
"""
from typing import Any
def compose_ordered_stories_to_markdown(
ordered: list[tuple[str, str]],
) -> str:
"""
:param ordered: (story_title, canonical_markdown) 已按阅读顺序排好
:return: 章节级 markdown每个故事为 ## 标题 + 正文,故事之间用 markdown 水平线 --- 分隔
(配图在 story 正文中,自然落在该故事块内、--- 之前)
"""
parts: list[str] = []
for title, md in ordered:
title = (title or "").strip() or "故事"
body = (md or "").strip()
parts.append(f"## {title}\n\n{body}" if body else f"## {title}")
return "\n\n---\n\n".join(parts)
from app.features.memoir.markdown_sanitize import sanitize_story_for_chapter_compose
def materialize_chapter_markdown_from_loaded_chapter(chapter: Any) -> str:
"""要求 chapter.story_links 已 eager-load且各 link.story 可用。"""
def _gather_title_body_pairs(chapter: Any) -> list[tuple[str, str]]:
links = sorted(
list(getattr(chapter, "story_links", None) or []),
key=lambda x: getattr(x, "order_index", 0),
@@ -36,4 +23,45 @@ def materialize_chapter_markdown_from_loaded_chapter(chapter: Any) -> str:
title = (getattr(st, "title", None) or "").strip()
body = (getattr(st, "canonical_markdown", None) or "").strip()
pairs.append((title, body))
return compose_ordered_stories_to_markdown(pairs)
return pairs
def compose_ordered_stories_to_markdown(ordered: list[tuple[str, str]]) -> str:
"""
:param ordered: (story_title, canonical_markdown) 已按阅读顺序排好title 仅用于清洗去重)
:return: 章节级 markdown仅各故事正文非空块之间用 \\n\\n---\\n\\n 分隔
"""
bodies: list[str] = []
for title, md in ordered:
raw = (md or "").strip()
if not raw:
continue
cleaned = sanitize_story_for_chapter_compose(raw, title)
if cleaned:
bodies.append(cleaned)
return "\n\n---\n\n".join(bodies)
def compose_ordered_stories_to_pdf_markdown(ordered: list[tuple[str, str]]) -> str:
"""PDF每故事 ## 标题 + 正文,块间 ---(标题来自元数据,不写回章节 canonical"""
parts: list[str] = []
for title, md in ordered:
t = (title or "").strip() or "故事"
raw = (md or "").strip()
if not raw:
continue
body = sanitize_story_for_chapter_compose(raw, title)
if not body:
continue
parts.append(f"## {t}\n\n{body}")
return "\n\n---\n\n".join(parts)
def materialize_chapter_markdown_from_loaded_chapter(chapter: Any) -> str:
"""要求 chapter.story_links 已 eager-load且各 link.story 可用。"""
return compose_ordered_stories_to_markdown(_gather_title_body_pairs(chapter))
def materialize_chapter_pdf_markdown_from_loaded_chapter(chapter: Any) -> str:
"""PDF 专用:含每段 ## 故事名。"""
return compose_ordered_stories_to_pdf_markdown(_gather_title_body_pairs(chapter))