53d9e003af
- 回忆录：事实边界补充允许清单；传记文体示例与 JSON 叙事要求对齐 - default 职业提示 occupation_context；cadre/military 退休语境 - GET 章节读路径零写入，prepare_chapter_read_view + markdown_for_response - 文本归一抽到 core/text_normalize；移除弃用 reply 策略与 recompose_chapters_for_story - ConversationService：WS 连接/用户段落/结束对话；对外错误固定文案 - 测试：HTTP 脱敏契约、章节读视图、occupation 与 background_voice
329 lines
13 KiB
Python
329 lines
13 KiB
Python
"""Memoir service — 回忆录编排（章节生成、状态流转）；通过 MemoryService 获取 evidence。"""
|
||
|
||
import asyncio
|
||
from typing import List, Optional
|
||
|
||
from fastapi import HTTPException
|
||
from sqlalchemy import select
|
||
from sqlalchemy.ext.asyncio import AsyncSession
|
||
from sqlalchemy.orm import joinedload
|
||
|
||
from app.core.logging import get_logger
|
||
from app.core.storage_purge import delete_object_storage_keys_best_effort
|
||
from app.features.memoir import repo
|
||
from app.features.memoir.asset_resolver import (
|
||
collect_asset_ids_for_chapter,
|
||
collect_asset_ids_for_chapters,
|
||
strip_image_placeholders,
|
||
)
|
||
from app.features.memoir.asset_urls import signed_urls_for_asset_ids
|
||
from app.features.memoir.chapter_markdown_compose import (
|
||
materialize_chapter_markdown_from_loaded_chapter,
|
||
)
|
||
from app.features.memoir.cover_eligibility import primary_chapter_memoir_image
|
||
from app.features.memoir.helpers import chapter_to_dict, chapter_to_list_dict
|
||
from app.features.memoir.memoir_images.settings import MemoirImageSettings
|
||
from app.features.memoir.models import Book, Chapter, ChapterStoryLink
|
||
from app.features.memoir.reading_segment_materialize import (
|
||
chapter_meets_minimum_display,
|
||
)
|
||
from app.features.memory.service import MemoryService
|
||
from app.ports.storage import ObjectStorage
|
||
|
||
logger = get_logger(__name__)
|
||
|
||
|
||
def prepare_chapter_read_view(chapter: Chapter) -> tuple[Chapter, str | None]:
|
||
"""
|
||
读路径：不写入 DB。返回 (chapter, markdown_for_response)。
|
||
None 表示序列化使用 ORM 上的 canonical_markdown。
|
||
"""
|
||
has_story_links = bool(getattr(chapter, "story_links", None))
|
||
has_snapshot = chapter.reading_segments_json is not None
|
||
dirty = chapter.markdown_compose_dirty is True
|
||
|
||
if not dirty:
|
||
return chapter, None
|
||
|
||
if has_snapshot:
|
||
return chapter, None
|
||
|
||
if has_story_links:
|
||
md = materialize_chapter_markdown_from_loaded_chapter(chapter)
|
||
m = (md or "").strip()
|
||
return chapter, m if m else None
|
||
|
||
return chapter, None
|
||
|
||
|
||
async def get_or_create_book(user_id: str, db: AsyncSession):
|
||
"""Get the user's current book or return None."""
|
||
return await repo.get_current_book(user_id, db)
|
||
|
||
|
||
class MemoirService:
|
||
def __init__(
|
||
self,
|
||
db: AsyncSession,
|
||
memory_service: Optional[MemoryService] = None,
|
||
*,
|
||
object_storage: ObjectStorage | None = None,
|
||
):
|
||
self._db = db
|
||
self._memory = memory_service
|
||
self._object_storage = object_storage
|
||
|
||
async def get_evidence(self, user_id: str, query: str, *, top_k: int = 10) -> dict:
|
||
"""通过 MemoryService → HybridRetriever 获取证据（含向量时与 Celery 的 FTS-only 路径不同）。"""
|
||
if self._memory is None:
|
||
return {
|
||
"relevant_chunks": [],
|
||
"relevant_summaries": [],
|
||
"relevant_facts": [],
|
||
"timeline_hints": [],
|
||
"relevant_stories": [],
|
||
}
|
||
bundle = await self._memory.retrieve(user_id, query, top_k=top_k)
|
||
return bundle.model_dump()
|
||
|
||
async def get_current_book(self, user_id: str) -> dict:
|
||
book = await repo.get_current_book(user_id, self._db)
|
||
if not book:
|
||
return {"message": "No book found"}
|
||
return {
|
||
"id": book.id,
|
||
"title": book.title,
|
||
"total_pages": book.total_pages,
|
||
"total_words": book.total_words,
|
||
"cover_image_url": book.cover_image_url,
|
||
"has_update": book.has_update,
|
||
"last_update_chapter_id": book.last_update_chapter_id,
|
||
}
|
||
|
||
async def clear_book_update(self, user_id: str) -> dict:
|
||
book = await repo.get_current_book(user_id, self._db)
|
||
if not book:
|
||
return {"status": "ok", "message": "No book found"}
|
||
book.has_update = False
|
||
await self._db.commit()
|
||
return {"status": "ok"}
|
||
|
||
async def update_book(self, book_id: str, user_id: str, title: str) -> dict:
|
||
book = await self._db.get(Book, book_id)
|
||
if not book:
|
||
raise HTTPException(status_code=404, detail="Book not found")
|
||
if book.user_id != user_id:
|
||
raise HTTPException(status_code=403, detail="无权更新此回忆录")
|
||
book.title = title
|
||
await self._db.commit()
|
||
await self._db.refresh(book)
|
||
return {
|
||
"id": book.id,
|
||
"title": book.title,
|
||
"total_pages": book.total_pages,
|
||
"total_words": book.total_words,
|
||
"cover_image_url": book.cover_image_url,
|
||
"has_update": book.has_update,
|
||
"last_update_chapter_id": book.last_update_chapter_id,
|
||
}
|
||
|
||
async def export_pdf(self, user_id: str, book_id: str) -> dict:
|
||
from app.features.memoir.pdf_service import pdf_service
|
||
|
||
book = await self._db.get(Book, book_id)
|
||
if not book:
|
||
raise HTTPException(status_code=404, detail="Book not found")
|
||
if book.user_id != user_id:
|
||
raise HTTPException(status_code=403, detail="无权导出此回忆录")
|
||
stmt = (
|
||
select(Chapter)
|
||
.where(Chapter.user_id == user_id, Chapter.is_active == True)
|
||
.options(
|
||
joinedload(Chapter.images),
|
||
joinedload(Chapter.story_links).joinedload(ChapterStoryLink.story),
|
||
)
|
||
.order_by(Chapter.order_index)
|
||
)
|
||
result = await self._db.execute(stmt)
|
||
chapters_raw = list(result.unique().scalars().all())
|
||
chapters: List[Chapter] = []
|
||
for ch in chapters_raw:
|
||
ch2, md_override = prepare_chapter_read_view(ch)
|
||
if chapter_meets_minimum_display(
|
||
ch2, canonical_markdown_override=md_override
|
||
):
|
||
chapters.append(ch2)
|
||
asset_ids = collect_asset_ids_for_chapters(chapters)
|
||
asset_map = await signed_urls_for_asset_ids(self._db, asset_ids)
|
||
pdf_bytes = await pdf_service.generate_pdf(
|
||
book, chapters, asset_url_map=asset_map
|
||
)
|
||
return {
|
||
"pdf_base64": pdf_bytes.decode("latin1"),
|
||
"filename": f"{book.title}.pdf",
|
||
}
|
||
|
||
async def get_chapters(
|
||
self, user_id: str, is_new: bool | None = None
|
||
) -> List[dict]:
|
||
chapters = await repo.get_chapters_for_memoir_list(
|
||
user_id, self._db, is_new_only=is_new
|
||
)
|
||
if not chapters:
|
||
return []
|
||
|
||
asset_ids: set[str] = set()
|
||
for ch in chapters:
|
||
asset_ids |= collect_asset_ids_for_chapter(ch)
|
||
asset_map = await signed_urls_for_asset_ids(self._db, asset_ids)
|
||
all_chapters: List[dict] = []
|
||
for ch in chapters:
|
||
ch, md_override = prepare_chapter_read_view(ch)
|
||
if not chapter_meets_minimum_display(
|
||
ch, canonical_markdown_override=md_override
|
||
):
|
||
continue
|
||
all_chapters.append(
|
||
chapter_to_list_dict(
|
||
ch,
|
||
asset_url_map=asset_map,
|
||
markdown_for_response=md_override,
|
||
)
|
||
)
|
||
return all_chapters
|
||
|
||
async def get_chapter(self, chapter_id: str, user_id: str) -> dict:
|
||
chapter = await repo.get_chapter_by_id(chapter_id, self._db)
|
||
if not chapter:
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
if chapter.user_id != user_id:
|
||
raise HTTPException(status_code=403, detail="无权访问此章节")
|
||
if not chapter.is_active:
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
chapter, md_override = prepare_chapter_read_view(chapter)
|
||
if not chapter_meets_minimum_display(
|
||
chapter, canonical_markdown_override=md_override
|
||
):
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
asset_map = await signed_urls_for_asset_ids(
|
||
self._db, collect_asset_ids_for_chapter(chapter)
|
||
)
|
||
return chapter_to_dict(
|
||
chapter,
|
||
asset_url_map=asset_map,
|
||
markdown_for_response=md_override,
|
||
)
|
||
|
||
async def disable_chapter(self, chapter_id: str, user_id: str) -> dict:
|
||
chapter = await repo.get_chapter_by_id(chapter_id, self._db)
|
||
if not chapter:
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
if chapter.user_id != user_id:
|
||
raise HTTPException(status_code=403, detail="无权操作此章节")
|
||
cos_keys = await repo.collect_cos_storage_keys_for_chapter(self._db, chapter)
|
||
chapter.is_active = False
|
||
await self._db.commit()
|
||
delete_object_storage_keys_best_effort(
|
||
self._object_storage,
|
||
cos_keys,
|
||
log_prefix=f"chapter_soft_delete id={chapter_id}",
|
||
)
|
||
return {"status": "ok", "message": "章节已清除"}
|
||
|
||
async def set_chapter_story_order(
|
||
self, chapter_id: str, user_id: str, story_ids: list[str]
|
||
) -> dict:
|
||
chapter = await self._db.get(Chapter, chapter_id)
|
||
if not chapter:
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
if chapter.user_id != user_id:
|
||
raise HTTPException(status_code=403, detail="无权操作此章节")
|
||
if not chapter.is_active:
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
try:
|
||
await repo.replace_chapter_story_links_async(
|
||
self._db,
|
||
chapter_id=chapter_id,
|
||
user_id=user_id,
|
||
story_ids=story_ids,
|
||
)
|
||
except ValueError as exc:
|
||
raise HTTPException(status_code=400, detail=str(exc)) from exc
|
||
|
||
ch = await repo.get_chapter_with_story_links_for_compose(chapter_id, self._db)
|
||
if not ch:
|
||
raise HTTPException(status_code=404, detail="Chapter not found")
|
||
if not ch.story_links:
|
||
md = ""
|
||
else:
|
||
md = materialize_chapter_markdown_from_loaded_chapter(ch)
|
||
await repo.append_chapter_compose_version_async(self._db, ch, md)
|
||
await self._db.commit()
|
||
return {"status": "ok", "chapter_id": chapter_id, "story_count": len(story_ids)}
|
||
|
||
async def get_memoir_state(self, user_id: str) -> dict:
|
||
from app.features.memoir.state_service import get_or_create_state
|
||
|
||
state = await get_or_create_state(user_id, self._db)
|
||
return state.model_dump()
|
||
|
||
async def get_next_question_context(self, user_id: str) -> dict:
|
||
from app.features.memoir.state_service import get_or_create_state
|
||
|
||
state = await get_or_create_state(user_id, self._db)
|
||
return {
|
||
"current_stage": state.current_stage,
|
||
"empty_slots": state.empty_slots_for_current_stage(),
|
||
"covered_stages": state.covered_stages,
|
||
}
|
||
|
||
async def check_and_trigger_cover_generation(self, user_id: str) -> dict:
|
||
"""
|
||
有正文、尚无 cover_asset、且封面 MemoirImage 未 completed 时，
|
||
派发 generate_chapter_cover（由 intent/asset 闭环完成）。
|
||
"""
|
||
from app.tasks.chapter_cover_enqueue import try_enqueue_generate_chapter_cover
|
||
|
||
img_settings = MemoirImageSettings.from_env()
|
||
if not img_settings.enabled:
|
||
return {"triggered": []}
|
||
|
||
chapters = await repo.get_chapters_for_memoir_list(
|
||
user_id, self._db, active_only=True, is_new_only=None
|
||
)
|
||
triggered: List[str] = []
|
||
for ch in chapters:
|
||
if not ch.category or not getattr(ch, "story_links", None):
|
||
continue
|
||
if getattr(ch, "cover_asset_id", None):
|
||
continue
|
||
md = (ch.canonical_markdown or "").strip()
|
||
body = strip_image_placeholders(md).strip() if md else ""
|
||
if not body:
|
||
continue
|
||
cover_rec = primary_chapter_memoir_image(ch)
|
||
if cover_rec and (cover_rec.status or "").strip() == "completed":
|
||
continue
|
||
enqueued = await asyncio.to_thread(
|
||
try_enqueue_generate_chapter_cover, ch.id, "http"
|
||
)
|
||
if enqueued:
|
||
triggered.append(ch.id)
|
||
logger.info("触发生成章节封面(asset): chapter={}", ch.id)
|
||
return {"triggered": triggered}
|
||
|
||
async def mark_memoir_read(self, user_id: str) -> dict:
|
||
stmt = select(Chapter).where(Chapter.user_id == user_id, Chapter.is_new == True)
|
||
result = await self._db.execute(stmt)
|
||
for chapter in result.scalars().all():
|
||
chapter.is_new = False
|
||
stmt_book = (
|
||
select(Book).where(Book.user_id == user_id).order_by(Book.updated_at.desc())
|
||
)
|
||
result_book = await self._db.execute(stmt_book)
|
||
book = result_book.scalar_one_or_none()
|
||
if book:
|
||
book.has_update = False
|
||
await self._db.commit()
|
||
return {"status": "ok"}
|