feat: 后台任务系统 + JSON容错解析 + SSE心跳保活 + 多项Bug修复

新功能:
- 大纲/章节生成改为服务端后台任务,支持断线续传
- 后台任务队列排队执行,按用户排队(同用户串行不同用户并发)
- 章节管理页面添加后台任务列表弹窗和进度面板
- 章节状态添加 pending(待处理)选项
- 集成json5容错解析器 + 上下文感知JSON修复
- SSE流式生成添加心跳保活,防止连接超时
- SSEPostClient添加credentials:include修复network error
- 每章最大伏笔数从2调整为5
- 添加大纲读区伏笔的功能

Bug修复:
- 修复AI生成JSON中未转义引号/中文标点/多对象属性值未合并
- 修复JSON非法转义字符清洗和中文引号处理
- 修复MCP插件TimeoutError/连接失败上下文清理
- MCP插件后台注册添加重试机制
- 续写模式添加缺失的mcp_references参数
- 修复Alembic迁移链分叉
- 使用torch CPU版本加速Docker构建
This commit is contained in:
未来
2026-04-29 08:31:07 +08:00
parent 1f80a58994
commit 2bd8b61e91
20 changed files with 2873 additions and 151 deletions
+445
View File
@@ -27,6 +27,7 @@ from app.models.analysis_task import AnalysisTask
from app.models.memory import PlotAnalysis, StoryMemory
from app.models.batch_generation_task import BatchGenerationTask
from app.models.regeneration_task import RegenerationTask
from app.models.background_task import BackgroundTask
from app.schemas.chapter import (
ChapterCreate,
ChapterUpdate,
@@ -1815,6 +1816,450 @@ async def generate_chapter_content_stream(
return create_sse_response(event_generator())
@router.post("/{chapter_id}/generate-background", summary="AI创作章节内容(后台任务)")
async def generate_chapter_content_background(
chapter_id: str,
request: Request,
generate_request: ChapterGenerateRequest = ChapterGenerateRequest(),
db: AsyncSession = Depends(get_db)
):
"""
创建后台任务来生成章节内容。
任务创建后立即返回task_id,前端通过 GET /api/tasks/{task_id} 轮询进度。
关闭浏览器不影响生成,生成完成后内容自动保存到数据库。
"""
user_id = getattr(request.state, 'user_id', None)
if not user_id:
raise HTTPException(status_code=401, detail="未登录")
# 验证章节存在
result = await db.execute(
select(Chapter).where(Chapter.id == chapter_id)
)
chapter = result.scalar_one_or_none()
if not chapter:
raise HTTPException(status_code=404, detail="章节不存在")
# 验证项目权限
project = await verify_project_access(chapter.project_id, user_id, db)
# 检查前置条件
can_generate, error_msg, _ = await check_prerequisites(db, chapter)
if not can_generate:
raise HTTPException(status_code=400, detail=error_msg)
# 创建后台任务
from app.services.background_task_service import background_task_service, TaskProgressTracker
task = await background_task_service.create_task(
user_id=user_id,
project_id=chapter.project_id,
task_type="chapter_generate",
task_input={
"chapter_id": chapter_id,
"style_id": generate_request.style_id,
"target_word_count": generate_request.target_word_count or 3000,
"enable_mcp": generate_request.enable_mcp,
"model": generate_request.model,
"narrative_perspective": generate_request.narrative_perspective,
},
db=db
)
# 后台执行的函数
async def _run_chapter_generation(task_id: str, bg_user_id: str):
from app.database import get_engine
from sqlalchemy.ext.asyncio import async_sessionmaker, AsyncSession as BgAsyncSession
engine = await get_engine(bg_user_id)
AsyncSessionLocal = async_sessionmaker(engine, class_=BgAsyncSession, expire_on_commit=False)
async with AsyncSessionLocal() as bg_db:
tracker = TaskProgressTracker(task_id, bg_user_id, "章节")
try:
await tracker.start()
# 获取AI服务
from app.api.settings import get_user_ai_service_from_db
bg_ai_service = await get_user_ai_service_from_db(bg_user_id, bg_db)
await _run_chapter_generation_bg(
task_input={
"chapter_id": chapter_id,
"style_id": generate_request.style_id,
"target_word_count": generate_request.target_word_count or 3000,
"enable_mcp": generate_request.enable_mcp,
"model": generate_request.model,
"narrative_perspective": generate_request.narrative_perspective,
},
db=bg_db,
ai_service=bg_ai_service,
tracker=tracker,
user_id=bg_user_id,
task_id=task_id,
)
except Exception as e:
logger.error(f"❌ 后台章节生成失败: {e}", exc_info=True)
await tracker.error(str(e))
await background_task_service.spawn_background_task(
task.id, user_id, _run_chapter_generation
)
return {
"task_id": task.id,
"task_type": "chapter_generate",
"status": "pending",
"message": "任务已创建,请通过 GET /api/tasks/{task_id} 查询进度"
}
async def _run_chapter_generation_bg(
task_input: dict,
db: AsyncSession,
ai_service: AIService,
tracker,
user_id: str,
task_id: str,
):
"""后台执行章节生成(不使用SSE,直接生成并保存)"""
from app.services.chapter_context_service import (
OneToManyContextBuilder,
OneToOneContextBuilder
)
chapter_id = task_input["chapter_id"]
style_id = task_input.get("style_id")
target_word_count = task_input.get("target_word_count", 3000)
custom_model = task_input.get("model")
temp_narrative_perspective = task_input.get("narrative_perspective")
write_lock = await get_db_write_lock(user_id)
# === 加载阶段 ===
await tracker.loading("加载章节信息...", 0.2)
chapter_result = await db.execute(
select(Chapter).where(Chapter.id == chapter_id)
)
current_chapter = chapter_result.scalar_one_or_none()
if not current_chapter:
await tracker.error("章节不存在")
return
await tracker.loading("加载项目信息...", 0.4)
project_result = await db.execute(
select(Project).where(Project.id == current_chapter.project_id)
)
project = project_result.scalar_one_or_none()
if not project:
await tracker.error("项目不存在")
return
outline_mode = project.outline_mode if project else 'one-to-many'
# 获取大纲
if current_chapter.outline_id:
outline_result = await db.execute(
select(Outline).where(Outline.id == current_chapter.outline_id)
)
else:
outline_result = await db.execute(
select(Outline)
.where(Outline.project_id == current_chapter.project_id)
.where(Outline.order_index == current_chapter.chapter_number)
)
outline = outline_result.scalar_one_or_none()
# 获取写作风格
style_content = ""
if style_id:
style_result = await db.execute(
select(WritingStyle).where(WritingStyle.id == style_id)
)
style = style_result.scalar_one_or_none()
if style and (style.user_id is None or style.user_id == user_id):
style_content = style.prompt_content or ""
# === 构建上下文 ===
if outline_mode == 'one-to-one':
context_builder = OneToOneContextBuilder(
memory_service=memory_service,
foreshadow_service=foreshadow_service
)
chapter_context = await context_builder.build(
chapter=current_chapter,
project=project,
outline=outline,
user_id=user_id,
db=db,
target_word_count=target_word_count
)
else:
context_builder = OneToManyContextBuilder(
memory_service=memory_service,
foreshadow_service=foreshadow_service
)
chapter_context = await context_builder.build(
chapter=current_chapter,
project=project,
outline=outline,
user_id=user_id,
db=db,
style_content=style_content,
target_word_count=target_word_count,
temp_narrative_perspective=temp_narrative_perspective
)
await tracker.loading("上下文构建完成", 0.8)
# 确定叙事人称
chapter_perspective = (
temp_narrative_perspective or
project.narrative_perspective or
'第三人称'
)
# === 准备提示词 ===
if outline_mode == 'one-to-one':
if chapter_context.continuation_point:
template = await PromptService.get_template("CHAPTER_GENERATION_ONE_TO_ONE_NEXT", user_id, db)
base_prompt = PromptService.format_prompt(
template,
project_title=project.title,
chapter_number=current_chapter.chapter_number,
chapter_title=current_chapter.title,
chapter_outline=chapter_context.chapter_outline,
target_word_count=target_word_count,
genre=project.genre or '未设定',
narrative_perspective=chapter_perspective,
previous_chapter_content=chapter_context.continuation_point,
previous_chapter_summary=chapter_context.previous_chapter_summary or '(无上一章摘要)',
characters_info=chapter_context.chapter_characters or '暂无角色信息',
chapter_careers=chapter_context.chapter_careers or '暂无职业信息',
foreshadow_reminders=chapter_context.foreshadow_reminders or '暂无需要关注的伏笔',
relevant_memories=chapter_context.relevant_memories or '暂无相关记忆'
)
else:
template = await PromptService.get_template("CHAPTER_GENERATION_ONE_TO_ONE", user_id, db)
base_prompt = PromptService.format_prompt(
template,
project_title=project.title,
chapter_number=current_chapter.chapter_number,
chapter_title=current_chapter.title,
chapter_outline=chapter_context.chapter_outline,
target_word_count=target_word_count,
genre=project.genre or '未设定',
narrative_perspective=chapter_perspective,
characters_info=chapter_context.chapter_characters or '暂无角色信息',
chapter_careers=chapter_context.chapter_careers or '暂无职业信息',
foreshadow_reminders=chapter_context.foreshadow_reminders or '暂无需要关注的伏笔',
relevant_memories=chapter_context.relevant_memories or '暂无相关记忆'
)
else:
if chapter_context.continuation_point:
previous_summary = chapter_context.previous_chapter_summary or "(无上一章摘要,请根据锚点续写)"
template = await PromptService.get_template("CHAPTER_GENERATION_ONE_TO_MANY_NEXT", user_id, db)
base_prompt = PromptService.format_prompt(
template,
project_title=project.title,
chapter_number=current_chapter.chapter_number,
chapter_title=current_chapter.title,
chapter_outline=chapter_context.chapter_outline,
target_word_count=target_word_count,
continuation_point=chapter_context.continuation_point,
genre=project.genre or '未设定',
narrative_perspective=chapter_perspective,
characters_info=chapter_context.chapter_characters or '暂无角色信息',
chapter_careers=chapter_context.chapter_careers or '暂无职业信息',
foreshadow_reminders=chapter_context.foreshadow_reminders or '暂无需要关注的伏笔',
previous_chapter_summary=previous_summary,
recent_chapters_context=chapter_context.recent_chapters_context or '',
relevant_memories=chapter_context.relevant_memories or ''
)
else:
template = await PromptService.get_template("CHAPTER_GENERATION_ONE_TO_MANY", user_id, db)
base_prompt = PromptService.format_prompt(
template,
project_title=project.title,
chapter_number=current_chapter.chapter_number,
chapter_title=current_chapter.title,
chapter_outline=chapter_context.chapter_outline,
target_word_count=target_word_count,
genre=project.genre or '未设定',
narrative_perspective=chapter_perspective,
characters_info=chapter_context.chapter_characters or '暂无角色信息',
chapter_careers=chapter_context.chapter_careers or '暂无职业信息',
foreshadow_reminders=chapter_context.foreshadow_reminders or '暂无需要关注的伏笔',
relevant_memories=chapter_context.relevant_memories or '暂无相关记忆'
)
# 应用写作风格
if style_content:
prompt = WritingStyleManager.apply_style_to_prompt(base_prompt, style_content)
else:
prompt = base_prompt
# === 准备阶段 ===
await tracker.preparing("准备AI提示词...")
system_prompt_with_style = None
if style_content:
system_prompt_with_style = f"""【🎨 写作风格要求 - 最高优先级】
{style_content}
⚠️ 请严格遵循上述写作风格要求进行创作,这是最重要的指令!
确保在整个章节创作过程中始终保持风格的一致性。"""
calculated_max_tokens = int(target_word_count * 3)
calculated_max_tokens = max(2000, min(calculated_max_tokens, 16000))
generate_kwargs = {
"prompt": prompt,
"system_prompt": system_prompt_with_style,
"tool_choice": "required",
"max_tokens": calculated_max_tokens
}
if custom_model:
generate_kwargs["model"] = custom_model
# === 生成阶段 ===
full_content = ""
chunk_count = 0
await tracker.generating(
current_chars=0,
estimated_total=target_word_count
)
async for chunk in ai_service.generate_text_stream(**generate_kwargs):
# 检查是否被取消
if chunk_count % 10 == 0 and await tracker.check_cancelled():
logger.info(f"🚫 后台章节生成被取消: {chapter_id}")
return
full_content += chunk
chunk_count += 1
# 每10个chunk更新一次进度
if chunk_count % 10 == 0:
await tracker.generating(
current_chars=len(full_content),
estimated_total=target_word_count,
message=f'正在创作中... 已生成 {len(full_content)}'
)
await asyncio.sleep(0)
# === 保存阶段 ===
await tracker.saving("正在保存章节...", 0.3)
async with write_lock:
# 重新获取章节(确保最新状态)
chapter_result = await db.execute(
select(Chapter).where(Chapter.id == chapter_id)
)
current_chapter = chapter_result.scalar_one_or_none()
if not current_chapter:
await tracker.error("保存时章节不存在")
return
old_word_count = current_chapter.word_count or 0
current_chapter.content = full_content
new_word_count = len(full_content)
current_chapter.word_count = new_word_count
current_chapter.status = "completed"
# 更新项目字数
project_result = await db.execute(
select(Project).where(Project.id == current_chapter.project_id)
)
project_obj = project_result.scalar_one_or_none()
if project_obj:
project_obj.current_words = (project_obj.current_words or 0) - old_word_count + new_word_count
# 记录生成历史
history = GenerationHistory(
project_id=current_chapter.project_id,
chapter_id=current_chapter.id,
prompt=f"创作章节: 第{current_chapter.chapter_number}{current_chapter.title}",
generated_content=full_content[:500] if len(full_content) > 500 else full_content,
model="default"
)
db.add(history)
await db.commit()
logger.info(f"✅ 后台创作章节 {chapter_id} 完成,共 {new_word_count}")
# 🔮 自动标记伏笔
try:
plant_result = await foreshadow_service.auto_plant_pending_foreshadows(
db=db,
project_id=current_chapter.project_id,
chapter_id=chapter_id,
chapter_number=current_chapter.chapter_number,
chapter_content=full_content
)
if plant_result.get('planted_count', 0) > 0:
logger.info(f"🔮 自动标记伏笔已埋入: {plant_result['planted_count']}")
except Exception as plant_error:
logger.warning(f"⚠️ 自动标记伏笔埋入失败: {str(plant_error)}")
# 创建分析任务
analysis_task = AnalysisTask(
chapter_id=chapter_id,
user_id=user_id,
project_id=current_chapter.project_id,
status='pending',
progress=0
)
db.add(analysis_task)
await db.commit()
await db.refresh(analysis_task)
logger.info(f"📋 后台生成:已创建分析任务: {analysis_task.id}")
await asyncio.sleep(0.05)
# 启动后台分析
asyncio.create_task(
analyze_chapter_background(
chapter_id=chapter_id,
user_id=user_id,
project_id=current_chapter.project_id,
task_id=analysis_task.id,
ai_service=ai_service
)
)
# === 完成 ===
await tracker.complete(f"创作完成!共 {new_word_count}")
# 更新任务结果
from app.services.background_task_service import background_task_service
from sqlalchemy.ext.asyncio import async_sessionmaker, AsyncSession as BgAsyncSession
from app.database import get_engine as bg_get_engine
try:
engine = await bg_get_engine(user_id)
AsyncSessionLocal = async_sessionmaker(engine, class_=BgAsyncSession, expire_on_commit=False)
async with AsyncSessionLocal() as result_db:
from sqlalchemy import update as sql_update
await result_db.execute(
sql_update(BackgroundTask)
.where(BackgroundTask.id == task_id)
.values(task_result={
"chapter_id": chapter_id,
"word_count": new_word_count,
"analysis_task_id": analysis_task.id
})
)
await result_db.commit()
except Exception as e:
logger.warning(f"⚠️ 更新任务结果失败: {e}")
def _build_analysis_task_status_payload(
chapter_id: str,
task: Optional[AnalysisTask],
+517
View File
@@ -1717,6 +1717,523 @@ async def continue_outline_generator(
yield await tracker.error(f"续写失败: {str(e)}")
@router.post("/generate", summary="AI生成/续写大纲(后台任务)")
async def generate_outline_task(
data: Dict[str, Any],
request: Request,
db: AsyncSession = Depends(get_db),
user_ai_service: AIService = Depends(get_user_ai_service)
):
"""
使用后台任务生成或续写小说大纲(不怕断连,关闭浏览器也继续运行)
返回task_id,前端通过 GET /api/tasks/{task_id} 轮询进度
支持模式:
- auto/new/continue(同 generate-stream
"""
from app.services.background_task_service import background_task_service, TaskProgressTracker
from app.database import get_engine
from sqlalchemy.ext.asyncio import async_sessionmaker, AsyncSession as NewAsyncSession
user_id = getattr(request.state, 'user_id', None)
project = await verify_project_access(data.get("project_id"), user_id, db)
# 判断模式
mode = data.get("mode", "auto")
existing_result = await db.execute(
select(Outline)
.where(Outline.project_id == data.get("project_id"))
.order_by(Outline.order_index)
)
existing_outlines = existing_result.scalars().all()
if mode == "auto":
mode = "continue" if existing_outlines else "new"
data["user_id"] = user_id
data["mode"] = mode
if mode == "continue" and not existing_outlines:
raise HTTPException(status_code=400, detail="续写模式需要已有大纲")
# 创建后台任务
task_type = "outline_new" if mode == "new" else "outline_continue"
task = await background_task_service.create_task(
user_id=user_id,
project_id=data.get("project_id"),
task_type=task_type,
task_input=data,
db=db
)
# 后台执行的函数
async def _run_outline_generation(task_id: str, user_id: str):
engine = await get_engine(user_id)
AsyncSessionLocal = async_sessionmaker(engine, class_=NewAsyncSession, expire_on_commit=False)
async with AsyncSessionLocal() as bg_db:
tracker = TaskProgressTracker(task_id, user_id, "大纲")
try:
await tracker.start()
# 获取AI服务(需要在后台创建新实例)
from app.api.settings import get_user_ai_service_from_db
bg_ai_service = await get_user_ai_service_from_db(user_id, bg_db)
if mode == "new":
await _run_new_outline_bg(data, bg_db, bg_ai_service, tracker)
else:
await _run_continue_outline_bg(data, bg_db, bg_ai_service, tracker, user_id)
except Exception as e:
logger.error(f"❌ 后台大纲生成失败: {e}", exc_info=True)
await tracker.error(str(e))
await background_task_service.spawn_background_task(
task.id, user_id, _run_outline_generation
)
return {
"task_id": task.id,
"task_type": task_type,
"status": "pending",
"message": "任务已创建,请通过 GET /api/tasks/{task_id} 查询进度"
}
async def _run_new_outline_bg(
data: Dict[str, Any],
db: AsyncSession,
user_ai_service: AIService,
tracker
):
"""后台执行全新大纲生成"""
from app.database import get_engine
from sqlalchemy.ext.asyncio import async_sessionmaker, AsyncSession as BgAsyncSession
project_id = data.get("project_id")
chapter_count = int(data.get("chapter_count", 10))
user_id_for_mcp = data.get("user_id")
await tracker.loading("加载项目信息...", 0.3)
result = await db.execute(select(Project).where(Project.id == project_id))
project = result.scalar_one_or_none()
if not project:
await tracker.error("项目不存在")
return
await tracker.loading(f"准备生成{chapter_count}章大纲...", 0.6)
characters_result = await db.execute(select(Character).where(Character.project_id == project_id))
characters = characters_result.scalars().all()
characters_info = _build_characters_info(characters)
if user_id_for_mcp:
user_ai_service.user_id = user_id_for_mcp
user_ai_service.db_session = db
await tracker.preparing("准备AI提示词...")
template = await PromptService.get_template("OUTLINE_CREATE", user_id_for_mcp, db)
prompt = PromptService.format_prompt(
template,
title=project.title,
theme=data.get("theme") or project.theme or "未设定",
genre=data.get("genre") or project.genre or "通用",
chapter_count=chapter_count,
narrative_perspective=data.get("narrative_perspective") or "第三人称",
time_period=project.world_time_period or "未设定",
location=project.world_location or "未设定",
atmosphere=project.world_atmosphere or "未设定",
rules=project.world_rules or "未设定",
characters_info=characters_info or "暂无角色信息",
requirements=data.get("requirements") or "",
mcp_references=""
)
model_param = data.get("model")
provider_param = data.get("provider")
estimated_total = chapter_count * 1000
accumulated_text = ""
chunk_count = 0
await tracker.generating(current_chars=0, estimated_total=estimated_total)
async for chunk in user_ai_service.generate_text_stream(
prompt=prompt, provider=provider_param, model=model_param
):
chunk_count += 1
accumulated_text += chunk
if chunk_count % 10 == 0:
if await tracker.check_cancelled():
await tracker.error("任务已取消")
return
await tracker.generating(
current_chars=len(accumulated_text),
estimated_total=estimated_total
)
await tracker.parsing("解析大纲数据...")
ai_content = accumulated_text
# 解析响应(带重试)
max_retries = 2
retry_count = 0
outline_data = None
while retry_count <= max_retries:
try:
outline_data = _parse_ai_response(ai_content, raise_on_error=True)
break
except JSONParseError:
retry_count += 1
if retry_count > max_retries:
outline_data = _parse_ai_response(ai_content, raise_on_error=False)
break
await tracker.retry(retry_count, max_retries, "JSON解析失败")
tracker.reset_generating_progress()
accumulated_text = ""
retry_prompt = prompt + "\n\n【重要提醒】请确保返回完整的JSON数组。"
async for chunk in user_ai_service.generate_text_stream(
prompt=retry_prompt, provider=provider_param, model=model_param
):
accumulated_text += chunk
ai_content = accumulated_text
# ✅ P0修复:先保存新数据,再删除旧数据
await tracker.saving("保存新大纲到数据库...", 0.2)
outlines = await _save_outlines(project_id, outline_data, db, start_index=1)
await db.commit() # 先提交新数据!
logger.info(f"✅ 新大纲已保存: {len(outlines)}")
# 新数据安全后,再清理旧数据
await tracker.saving("清理旧数据...", 0.6)
try:
from sqlalchemy import delete as sql_delete
# 获取旧大纲(不包括刚保存的)
new_outline_ids = [o.id for o in outlines]
old_outlines_result = await db.execute(
select(Outline).where(
Outline.project_id == project_id,
~Outline.id.in_(new_outline_ids)
)
)
old_outlines = old_outlines_result.scalars().all()
if old_outlines:
old_outline_ids = [o.id for o in old_outlines]
# 清理旧章节
old_chapters_result = await db.execute(
select(Chapter).where(
Chapter.project_id == project_id,
~Chapter.id.in_([ch.id for ch in await db.execute(
select(Chapter).where(Chapter.outline_id.in_(new_outline_ids) if new_outline_ids else False)
).scalars().all()] if new_outline_ids else [])
)
)
# 简化:删除不属于新大纲的旧章节
# 先获取新大纲对应的章节(one-to-one模式下通过chapter_number匹配)
new_order_indexes = [o.order_index for o in outlines]
if project.outline_mode == 'one-to-one':
old_chapters_result = await db.execute(
select(Chapter).where(
Chapter.project_id == project_id,
~Chapter.chapter_number.in_(new_order_indexes)
)
)
else:
old_chapters_result = await db.execute(
select(Chapter).where(
Chapter.project_id == project_id,
Chapter.outline_id.in_(old_outline_ids)
)
)
old_chapters = old_chapters_result.scalars().all()
deleted_word_count = sum(ch.word_count or 0 for ch in old_chapters)
# 清理伏笔和记忆
for ch in old_chapters:
try:
await memory_service.delete_chapter_memories(
user_id=user_id_for_mcp, project_id=project_id, chapter_id=ch.id
)
except Exception:
pass
try:
await foreshadow_service.delete_chapter_foreshadows(
db=db, project_id=project_id, chapter_id=ch.id, only_analysis_source=True
)
except Exception:
pass
# 删除旧章节
if project.outline_mode == 'one-to-one':
await db.execute(
sql_delete(Chapter).where(
Chapter.project_id == project_id,
~Chapter.chapter_number.in_(new_order_indexes)
)
)
else:
await db.execute(
sql_delete(Chapter).where(Chapter.outline_id.in_(old_outline_ids))
)
if deleted_word_count > 0:
project.current_words = max(0, project.current_words - deleted_word_count)
# 清理伏笔
try:
await foreshadow_service.clear_project_foreshadows_for_reset(db, project_id)
except Exception:
pass
# 清理分析
try:
from app.models.memory import PlotAnalysis
await db.execute(sql_delete(PlotAnalysis).where(PlotAnalysis.project_id == project_id))
except Exception:
pass
# 删除旧大纲
await db.execute(
sql_delete(Outline).where(Outline.id.in_(old_outline_ids))
)
await db.commit()
logger.info(f"✅ 旧数据清理完成: 删除 {len(old_outlines)} 个旧大纲, {len(old_chapters)} 个旧章节")
except Exception as e:
logger.error(f"❌ 清理旧数据失败(新数据已安全保存): {e}")
# 新数据已保存,旧数据清理失败不影响
# 角色校验
await tracker.saving("🎭 校验角色信息...", 0.7)
try:
await _check_and_create_missing_characters_from_outlines(
outline_data=outline_data, project_id=project_id, db=db,
user_ai_service=user_ai_service, user_id=data.get("user_id"),
enable_mcp=data.get("enable_mcp", True), tracker=tracker
)
except Exception:
pass
# 组织校验
try:
await _check_and_create_missing_organizations_from_outlines(
outline_data=outline_data, project_id=project_id, db=db,
user_ai_service=user_ai_service, user_id=data.get("user_id"),
enable_mcp=data.get("enable_mcp", True), tracker=tracker
)
except Exception:
pass
# 保存结果到任务记录
result_data = {
"message": f"成功生成{len(outlines)}章大纲",
"total_chapters": len(outlines),
"outline_ids": [o.id for o in outlines]
}
# 更新任务结果
from app.models.background_task import BackgroundTask
task_result = await db.execute(select(BackgroundTask).where(BackgroundTask.id == tracker.task_id))
bg_task = task_result.scalar_one_or_none()
if bg_task:
bg_task.task_result = result_data
await db.commit()
await tracker.complete(f"成功生成{len(outlines)}章大纲")
logger.info(f"✅ 后台大纲生成完成: {len(outlines)}")
async def _run_continue_outline_bg(
data: Dict[str, Any],
db: AsyncSession,
user_ai_service: AIService,
tracker,
user_id: str
):
"""后台执行大纲续写"""
project_id = data.get("project_id")
total_chapters = int(data.get("chapter_count", 5))
await tracker.loading("加载项目信息...", 0.2)
result = await db.execute(select(Project).where(Project.id == project_id))
project = result.scalar_one_or_none()
if not project:
await tracker.error("项目不存在")
return
existing_result = await db.execute(
select(Outline).where(Outline.project_id == project_id).order_by(Outline.order_index)
)
existing_outlines = existing_result.scalars().all()
if not existing_outlines:
await tracker.error("续写模式需要已有大纲")
return
last_chapter_number = existing_outlines[-1].order_index
characters_result = await db.execute(select(Character).where(Character.project_id == project_id))
characters = characters_result.scalars().all()
batch_size = 5
total_batches = (total_chapters + batch_size - 1) // batch_size
all_new_outlines = []
current_start_chapter = last_chapter_number + 1
stage_instructions = {
"development": "继续展开情节,深化角色关系",
"climax": "进入故事高潮,矛盾激化",
"ending": "解决主要冲突,给出结局"
}
stage_instruction = stage_instructions.get(data.get("plot_stage", "development"), "")
for batch_num in range(total_batches):
if await tracker.check_cancelled():
await tracker.error("任务已取消")
return
remaining = total_chapters - len(all_new_outlines)
current_batch_size = min(batch_size, remaining)
tracker.reset_generating_progress()
await tracker.generating(
message=f"📝 第{batch_num + 1}/{total_batches}批: 生成第{current_start_chapter}-{current_start_chapter + current_batch_size - 1}"
)
latest_result = await db.execute(
select(Outline).where(Outline.project_id == project_id).order_by(Outline.order_index)
)
latest_outlines = latest_result.scalars().all()
context = await _build_outline_continue_context(
project=project, latest_outlines=latest_outlines, characters=characters,
chapter_count=current_batch_size,
plot_stage=data.get("plot_stage", "development"),
story_direction=data.get("story_direction", "自然延续"),
requirements=data.get("requirements", ""), db=db
)
user_ai_service.user_id = user_id
user_ai_service.db_session = db
# 获取伏笔提醒
foreshadow_reminders_text = "暂无需要关注的伏笔"
try:
foreshadow_context = await foreshadow_service.build_chapter_context(
db=db, project_id=project_id, chapter_number=current_start_chapter,
include_pending=False, include_overdue=True, lookahead=10
)
if foreshadow_context and foreshadow_context.get("context_text"):
foreshadow_reminders_text = foreshadow_context["context_text"]
except Exception:
pass
template = await PromptService.get_template("OUTLINE_CONTINUE", user_id, db)
prompt = PromptService.format_prompt(
template,
title=project.title, theme=project.theme or "未设定",
genre=project.genre or "通用",
narrative_perspective=project.narrative_perspective or "第三人称",
time_period=project.world_time_period or "未设定",
location=project.world_location or "未设定",
atmosphere=project.world_atmosphere or "未设定",
rules=project.world_rules or "未设定",
recent_outlines=context['recent_outlines'],
characters_info=context['characters_info'],
foreshadow_reminders=foreshadow_reminders_text,
chapter_count=current_batch_size,
start_chapter=current_start_chapter,
end_chapter=current_start_chapter + current_batch_size - 1,
current_chapter_count=len(latest_outlines),
plot_stage_instruction=stage_instruction,
story_direction=data.get("story_direction", "自然延续"),
requirements=data.get("requirements", ""),
mcp_references=""
)
accumulated_text = ""
chunk_count = 0
estimated_chars = current_batch_size * 1000
async for chunk in user_ai_service.generate_text_stream(
prompt=prompt, provider=data.get("provider"), model=data.get("model")
):
chunk_count += 1
accumulated_text += chunk
if chunk_count % 10 == 0:
await tracker.generating(
current_chars=len(accumulated_text), estimated_total=estimated_chars,
message=f"📝 第{batch_num + 1}/{total_batches}批生成中..."
)
await tracker.parsing(f"解析第{batch_num + 1}批数据...")
# 解析
max_retries = 2
retry_count = 0
outline_data = None
while retry_count <= max_retries:
try:
outline_data = _parse_ai_response(accumulated_text, raise_on_error=True)
break
except JSONParseError:
retry_count += 1
if retry_count > max_retries:
outline_data = _parse_ai_response(accumulated_text, raise_on_error=False)
break
await tracker.retry(retry_count, max_retries, "JSON解析失败")
tracker.reset_generating_progress()
accumulated_text = ""
retry_prompt = prompt + "\n\n【重要提醒】请确保返回完整的JSON数组。"
async for chunk in user_ai_service.generate_text_stream(
prompt=retry_prompt, provider=data.get("provider"), model=data.get("model")
):
accumulated_text += chunk
# 保存当前批次
await tracker.saving(f"保存第{batch_num + 1}批大纲...", 0.5)
batch_outlines = await _save_outlines(
project_id, outline_data, db, start_index=current_start_chapter
)
await db.commit()
all_new_outlines.extend(batch_outlines)
current_start_chapter += current_batch_size
# 角色校验
try:
await _check_and_create_missing_characters_from_outlines(
outline_data=outline_data, project_id=project_id, db=db,
user_ai_service=user_ai_service, user_id=user_id,
enable_mcp=data.get("enable_mcp", True), tracker=tracker
)
await db.commit()
except Exception:
pass
# 保存结果
result_data = {
"message": f"成功续写{len(all_new_outlines)}章大纲",
"total_chapters": len(all_new_outlines),
"outline_ids": [o.id for o in all_new_outlines]
}
from app.models.background_task import BackgroundTask
task_result = await db.execute(select(BackgroundTask).where(BackgroundTask.id == tracker.task_id))
bg_task = task_result.scalar_one_or_none()
if bg_task:
bg_task.task_result = result_data
await db.commit()
await tracker.complete(f"成功续写{len(all_new_outlines)}章大纲")
logger.info(f"✅ 后台大纲续写完成: {len(all_new_outlines)}")
@router.post("/generate-stream", summary="AI生成/续写大纲(SSE流式)")
async def generate_outline_stream(
data: Dict[str, Any],
+38
View File
@@ -160,6 +160,44 @@ async def get_user_ai_service(
)
async def get_user_ai_service_from_db(user_id: str, db: AsyncSession) -> AIService:
"""
从数据库直接创建用户AI服务实例用于后台任务不依赖FastAPI的Depends
"""
from app.models.mcp_plugin import MCPPlugin
result = await db.execute(
select(Settings).where(Settings.user_id == user_id)
)
settings = result.scalar_one_or_none()
if not settings:
env_defaults = read_env_defaults()
settings = Settings(user_id=user_id, **env_defaults)
db.add(settings)
await db.commit()
await db.refresh(settings)
mcp_result = await db.execute(
select(MCPPlugin).where(MCPPlugin.user_id == user_id)
)
mcp_plugins = mcp_result.scalars().all()
enable_mcp = any(plugin.enabled for plugin in mcp_plugins) if mcp_plugins else False
return create_user_ai_service_with_mcp(
api_provider=settings.api_provider,
api_key=settings.api_key,
api_base_url=settings.api_base_url or "",
model_name=settings.llm_model,
temperature=settings.temperature,
max_tokens=settings.max_tokens,
user_id=user_id,
db_session=db,
system_prompt=settings.system_prompt,
enable_mcp=enable_mcp,
)
@router.get("", response_model=SettingsResponse)
async def get_settings(
user: User = Depends(require_login),
+122
View File
@@ -0,0 +1,122 @@
"""后台任务API - 查询状态、取消任务"""
from fastapi import APIRouter, Depends, HTTPException, Request
from sqlalchemy.ext.asyncio import AsyncSession
from typing import Optional
from app.database import get_db
from app.models.background_task import BackgroundTask
from app.services.background_task_service import background_task_service
from app.logger import get_logger
router = APIRouter(prefix="/tasks", tags=["后台任务"])
logger = get_logger(__name__)
@router.get("/{task_id}", summary="获取任务状态")
async def get_task_status(
task_id: str,
request: Request,
db: AsyncSession = Depends(get_db)
):
"""获取后台任务的状态和进度"""
user_id = getattr(request.state, 'user_id', None)
if not user_id:
raise HTTPException(status_code=401, detail="未登录")
task = await background_task_service.get_task(task_id, user_id, db)
if not task:
raise HTTPException(status_code=404, detail="任务不存在")
return {
"id": task.id,
"task_type": task.task_type,
"project_id": task.project_id,
"status": task.status,
"progress": task.progress,
"status_message": task.status_message,
"progress_details": task.progress_details,
"error_message": task.error_message,
"task_result": task.task_result,
"retry_count": task.retry_count,
"cancel_requested": task.cancel_requested,
"created_at": task.created_at.isoformat() if task.created_at else None,
"started_at": task.started_at.isoformat() if task.started_at else None,
"completed_at": task.completed_at.isoformat() if task.completed_at else None,
"updated_at": task.updated_at.isoformat() if task.updated_at else None,
}
@router.get("", summary="获取任务列表")
async def get_tasks(
project_id: str,
request: Request,
task_type: Optional[str] = None,
limit: int = 20,
db: AsyncSession = Depends(get_db)
):
"""获取项目的后台任务列表"""
user_id = getattr(request.state, 'user_id', None)
if not user_id:
raise HTTPException(status_code=401, detail="未登录")
tasks = await background_task_service.get_project_tasks(
project_id, user_id, db, task_type=task_type, limit=limit
)
return {
"items": [
{
"id": t.id,
"task_type": t.task_type,
"status": t.status,
"progress": t.progress,
"status_message": t.status_message,
"progress_details": t.progress_details,
"error_message": t.error_message,
"created_at": t.created_at.isoformat() if t.created_at else None,
"completed_at": t.completed_at.isoformat() if t.completed_at else None,
}
for t in tasks
]
}
@router.post("/{task_id}/cancel", summary="取消任务")
async def cancel_task(
task_id: str,
request: Request,
db: AsyncSession = Depends(get_db)
):
"""请求取消后台任务"""
user_id = getattr(request.state, 'user_id', None)
if not user_id:
raise HTTPException(status_code=401, detail="未登录")
success = await background_task_service.cancel_task(task_id, user_id, db)
if not success:
raise HTTPException(status_code=400, detail="无法取消任务(不存在或已完成)")
return {"message": "任务已取消", "task_id": task_id}
@router.delete("/{task_id}", summary="删除任务记录")
async def delete_task(
task_id: str,
request: Request,
db: AsyncSession = Depends(get_db)
):
"""删除已完成/失败的任务记录"""
user_id = getattr(request.state, 'user_id', None)
if not user_id:
raise HTTPException(status_code=401, detail="未登录")
task = await background_task_service.get_task(task_id, user_id, db)
if not task:
raise HTTPException(status_code=404, detail="任务不存在")
if task.status in ("pending", "running"):
raise HTTPException(status_code=400, detail="无法删除进行中的任务,请先取消")
await db.delete(task)
await db.commit()
return {"message": "任务记录已删除"}
+2 -1
View File
@@ -21,7 +21,8 @@ from app.models import (
Settings, WritingStyle, ProjectDefaultStyle,
RelationshipType, CharacterRelationship, Organization, OrganizationMember,
StoryMemory, PlotAnalysis, AnalysisTask, BatchGenerationTask,
RegenerationTask, Career, CharacterCareer, User, MCPPlugin, PromptTemplate
RegenerationTask, Career, CharacterCareer, User, MCPPlugin, PromptTemplate,
BackgroundTask
)
# 引擎缓存:每个用户一个引擎
+17 -2
View File
@@ -29,7 +29,21 @@ async def lifespan(app: FastAPI):
"""应用生命周期管理"""
# 注册MCP状态同步服务
register_status_sync()
# 安全保障:确保后台任务表存在(兼容未执行Alembic迁移的旧部署)
try:
from app.database import get_engine
from app.models.background_task import BackgroundTask
_startup_engine = await get_engine("system")
async with _startup_engine.begin() as conn:
# 仅创建 background_tasks 表(如果不存在),不影响其他表
await conn.run_sync(
lambda sync_conn: BackgroundTask.__table__.create(sync_conn, checkfirst=True)
)
logger.info("后台任务表检查完成")
except Exception as e:
logger.warning(f"后台任务表检查失败(不影响启动): {e}")
logger.info("应用启动完成")
yield
@@ -133,7 +147,7 @@ from app.api import (
auth, users, settings, writing_styles, memories,
mcp_plugins, admin, inspiration, prompt_templates,
changelog, careers, foreshadows, prompt_workshop, book_import,
project_covers
project_covers, tasks
)
app.include_router(auth.router, prefix="/api")
@@ -159,6 +173,7 @@ app.include_router(prompt_templates.router, prefix="/api") # 提示词模板管
app.include_router(changelog.router, prefix="/api") # 更新日志API
app.include_router(prompt_workshop.router, prefix="/api") # 提示词工坊API
app.include_router(book_import.router, prefix="/api") # 拆书导入API
app.include_router(tasks.router, prefix="/api") # 后台任务API
static_dir = Path(__file__).parent.parent / "static"
generated_assets_root_dir = Path(__file__).parent.parent / "storage"
+4 -2
View File
@@ -18,6 +18,7 @@ from app.models.career import Career, CharacterCareer
from app.models.prompt_template import PromptTemplate
from app.models.foreshadow import Foreshadow
from app.models.prompt_workshop import PromptWorkshopItem, PromptSubmission, PromptWorkshopLike
from app.models.background_task import BackgroundTask
__all__ = [
"Project",
@@ -46,5 +47,6 @@ __all__ = [
"Foreshadow",
"PromptWorkshopItem",
"PromptSubmission",
"PromptWorkshopLike"
]
"PromptWorkshopLike",
"BackgroundTask"
]
+46
View File
@@ -0,0 +1,46 @@
"""后台任务数据模型 - 用于长时间运行的AI生成任务"""
from sqlalchemy import Column, String, Integer, DateTime, Boolean, JSON, Text
from sqlalchemy.sql import func
from app.database import Base
import uuid
class BackgroundTask(Base):
"""后台任务表 - 追踪所有长时间运行的生成任务"""
__tablename__ = "background_tasks"
id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
user_id = Column(String(100), nullable=False, index=True, comment="用户ID")
project_id = Column(String(36), nullable=False, index=True, comment="项目ID")
# 任务类型
task_type = Column(String(50), nullable=False, comment="任务类型: outline_new/outline_continue/outline_expand/chapter_generate/chapter_batch/wizard")
# 任务状态
status = Column(String(20), default="pending", comment="任务状态: pending/running/completed/failed/cancelled")
progress = Column(Integer, default=0, comment="进度百分比(0-100)")
status_message = Column(String(500), comment="当前状态消息")
# 任务输入/输出
task_input = Column(JSON, comment="任务输入参数(JSON)")
task_result = Column(JSON, comment="任务结果(JSON)")
error_message = Column(Text, comment="错误信息")
# 进度详情(用于前端展示实时进度)
progress_details = Column(JSON, comment="进度详情: {stage, message, word_count, etc.}")
# 取消支持
cancel_requested = Column(Boolean, default=False, comment="是否请求取消")
# 重试信息
retry_count = Column(Integer, default=0, comment="已重试次数")
max_retries = Column(Integer, default=3, comment="最大重试次数")
# 时间记录
created_at = Column(DateTime, server_default=func.now(), comment="创建时间")
started_at = Column(DateTime, comment="开始时间")
completed_at = Column(DateTime, comment="完成时间")
updated_at = Column(DateTime, server_default=func.now(), onupdate=func.now(), comment="更新时间")
def __repr__(self):
return f"<BackgroundTask(id={self.id[:8]}, type={self.task_type}, status={self.status}, progress={self.progress}%)>"
+1
View File
@@ -138,6 +138,7 @@ class BatchGenerateRequest(BaseModel):
enable_mcp: bool = Field(True, description="是否启用MCP工具增强(搜索参考资料)")
max_retries: int = Field(3, description="每个章节的最大重试次数", ge=0, le=5)
model: Optional[str] = Field(None, description="指定使用的AI模型,不提供则使用用户默认模型")
narrative_perspective: Optional[str] = Field(None, description="临时指定叙事人称,不提供则使用项目默认")
class BatchGenerateResponse(BaseModel):
@@ -0,0 +1,387 @@
"""后台任务管理服务 - 管理长时间运行的AI生成任务"""
import asyncio
from datetime import datetime
from typing import Dict, Any, Optional, Callable, Awaitable
from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker
from sqlalchemy import select, update
from app.database import get_engine
from app.models.background_task import BackgroundTask
from app.logger import get_logger
logger = get_logger(__name__)
class TaskProgressTracker:
"""后台任务进度追踪器(替代SSE的WizardProgressTracker"""
def __init__(self, task_id: str, user_id: str, task_name: str = "任务"):
self.task_id = task_id
self.user_id = user_id
self.task_name = task_name
self.current_progress = 0
self._last_generating_progress = 20
async def _update_task(self, **kwargs):
"""更新任务状态到数据库"""
try:
engine = await get_engine(self.user_id)
AsyncSessionLocal = async_sessionmaker(
engine, class_=AsyncSession, expire_on_commit=False
)
async with AsyncSessionLocal() as session:
result = await session.execute(
select(BackgroundTask).where(BackgroundTask.id == self.task_id)
)
task = result.scalar_one_or_none()
if task:
for key, value in kwargs.items():
setattr(task, key, value)
task.updated_at = datetime.now()
await session.commit()
except Exception as e:
logger.error(f"❌ 更新任务进度失败: {e}")
async def start(self, message: str = None):
self.current_progress = 0
msg = message or f"开始生成{self.task_name}..."
await self._update_task(
status="running", progress=0, status_message=msg,
started_at=datetime.now(),
progress_details={"stage": "init", "message": msg}
)
async def loading(self, message: str = None, sub_progress: float = 0.5):
progress = 5 + int(10 * sub_progress)
self.current_progress = progress
msg = message or "加载数据中..."
await self._update_task(
progress=progress, status_message=msg,
progress_details={"stage": "loading", "message": msg}
)
async def preparing(self, message: str = None):
self.current_progress = 17
msg = message or "准备AI提示词..."
await self._update_task(
progress=17, status_message=msg,
progress_details={"stage": "preparing", "message": msg}
)
async def generating(self, current_chars: int = 0, estimated_total: int = 5000,
message: str = None, retry_count: int = 0, max_retries: int = 3):
sub_progress = min(current_chars / max(estimated_total, 1), 1.0)
progress = 20 + int(65 * sub_progress)
if progress < self._last_generating_progress:
progress = self._last_generating_progress
else:
self._last_generating_progress = progress
self.current_progress = progress
retry_suffix = f" (重试 {retry_count}/{max_retries})" if retry_count > 0 else ""
msg = message or f"生成{self.task_name}中... ({current_chars}字符){retry_suffix}"
await self._update_task(
progress=progress, status_message=msg,
progress_details={"stage": "generating", "message": msg, "current_chars": current_chars}
)
async def parsing(self, message: str = None):
self.current_progress = 88
msg = message or f"解析{self.task_name}数据..."
await self._update_task(
progress=88, status_message=msg,
progress_details={"stage": "parsing", "message": msg}
)
async def saving(self, message: str = None, sub_progress: float = 0.5):
progress = 92 + int(6 * sub_progress)
self.current_progress = progress
msg = message or f"保存{self.task_name}到数据库..."
await self._update_task(
progress=progress, status_message=msg,
progress_details={"stage": "saving", "message": msg}
)
async def complete(self, message: str = None):
self.current_progress = 100
msg = message or f"{self.task_name}生成完成!"
await self._update_task(
status="completed", progress=100, status_message=msg,
completed_at=datetime.now(),
progress_details={"stage": "complete", "message": msg}
)
async def error(self, error_message: str):
await self._update_task(
status="failed", error_message=error_message,
status_message=f"失败: {error_message}",
completed_at=datetime.now(),
progress_details={"stage": "error", "message": error_message}
)
async def warning(self, message: str):
await self._update_task(
status_message=f"⚠️ {message}",
progress_details={"stage": "warning", "message": message}
)
async def retry(self, retry_count: int, max_retries: int, reason: str = "准备重试"):
msg = f"⚠️ {reason}... ({retry_count}/{max_retries})"
await self._update_task(
status_message=msg, retry_count=retry_count,
progress_details={"stage": "retry", "message": msg, "retry_count": retry_count}
)
def reset_generating_progress(self):
self._last_generating_progress = 20
async def check_cancelled(self) -> bool:
"""检查任务是否被取消"""
try:
engine = await get_engine(self.user_id)
AsyncSessionLocal = async_sessionmaker(
engine, class_=AsyncSession, expire_on_commit=False
)
async with AsyncSessionLocal() as session:
result = await session.execute(
select(BackgroundTask.cancel_requested)
.where(BackgroundTask.id == self.task_id)
)
cancelled = result.scalar_one_or_none()
return bool(cancelled)
except Exception:
return False
class BackgroundTaskService:
"""后台任务管理服务(按用户排队:同用户任务逐个执行,不同用户可并发)"""
def __init__(self):
self._user_queues: Dict[str, asyncio.Queue] = {} # user_id -> Queue
self._user_workers: Dict[str, bool] = {} # user_id -> worker是否运行中
def _ensure_user_queue(self, user_id: str) -> asyncio.Queue:
"""确保指定用户的队列已初始化"""
if user_id not in self._user_queues:
self._user_queues[user_id] = asyncio.Queue()
return self._user_queues[user_id]
async def _start_user_worker(self, user_id: str):
"""启动指定用户的工作协程"""
if self._user_workers.get(user_id, False):
return
self._user_workers[user_id] = True
asyncio.create_task(self._user_worker_loop(user_id))
logger.info(f"📋 用户 {user_id[:8]} 的任务队列工作协程已启动")
async def _user_worker_loop(self, user_id: str):
"""从指定用户的队列中逐个取出任务并执行"""
queue = self._user_queues[user_id]
try:
while True:
try:
task_item = await queue.get()
task_id = task_item["task_id"]
task_func = task_item["task_func"]
args = task_item["args"]
kwargs = task_item["kwargs"]
logger.info(f"🔄 [用户{user_id[:8]}] 队列开始执行任务: {task_id[:8]} (队列剩余: {queue.qsize()})")
try:
await task_func(task_id, args["user_id"], *args["extra_args"], **kwargs)
except Exception as e:
logger.error(f"❌ 后台任务 {task_id[:8]} 异常: {e}", exc_info=True)
# 确保任务状态更新为失败
try:
engine = await get_engine(user_id)
AsyncSessionLocal = async_sessionmaker(
engine, class_=AsyncSession, expire_on_commit=False
)
async with AsyncSessionLocal() as session:
result = await session.execute(
select(BackgroundTask).where(BackgroundTask.id == task_id)
)
task = result.scalar_one_or_none()
if task and task.status == "running":
task.status = "failed"
task.error_message = str(e)
task.status_message = f"任务失败: {str(e)}"
task.completed_at = datetime.now()
await session.commit()
except Exception as update_err:
logger.error(f"❌ 更新失败任务状态失败: {update_err}")
finally:
queue.task_done()
logger.info(f"✅ [用户{user_id[:8]}] 队列任务完成: {task_id[:8]} (队列剩余: {queue.qsize()})")
except Exception as e:
logger.error(f"❌ [用户{user_id[:8]}] 队列工作循环异常: {e}", exc_info=True)
finally:
# 工作协程退出时清理标记
self._user_workers.pop(user_id, None)
logger.info(f"📋 用户 {user_id[:8]} 的工作协程已退出")
@staticmethod
async def create_task(
user_id: str,
project_id: str,
task_type: str,
task_input: Dict[str, Any] = None,
db: AsyncSession = None
) -> BackgroundTask:
"""创建后台任务记录"""
task = BackgroundTask(
user_id=user_id,
project_id=project_id,
task_type=task_type,
task_input=task_input or {},
status="pending",
progress=0,
status_message="任务已创建,等待执行..."
)
db.add(task)
await db.commit()
await db.refresh(task)
logger.info(f"📋 创建后台任务: {task.id[:8]} type={task_type} project={project_id[:8]}")
return task
@staticmethod
async def get_task(task_id: str, user_id: str, db: AsyncSession) -> Optional[BackgroundTask]:
"""获取任务详情"""
result = await db.execute(
select(BackgroundTask).where(
BackgroundTask.id == task_id,
BackgroundTask.user_id == user_id
)
)
return result.scalar_one_or_none()
@staticmethod
async def get_project_tasks(
project_id: str, user_id: str, db: AsyncSession,
task_type: str = None, limit: int = 20
) -> list:
"""获取项目的任务列表"""
query = (
select(BackgroundTask)
.where(
BackgroundTask.project_id == project_id,
BackgroundTask.user_id == user_id
)
.order_by(BackgroundTask.created_at.desc())
)
if task_type:
query = query.where(BackgroundTask.task_type == task_type)
query = query.limit(limit)
result = await db.execute(query)
return result.scalars().all()
@staticmethod
async def cancel_task(task_id: str, user_id: str, db: AsyncSession) -> bool:
"""请求取消任务"""
result = await db.execute(
select(BackgroundTask).where(
BackgroundTask.id == task_id,
BackgroundTask.user_id == user_id
)
)
task = result.scalar_one_or_none()
if not task:
return False
if task.status not in ("pending", "running"):
return False
task.cancel_requested = True
task.status = "cancelled"
task.status_message = "任务已取消"
task.completed_at = datetime.now()
await db.commit()
logger.info(f"🚫 取消任务: {task_id[:8]}")
return True
@staticmethod
async def cleanup_old_tasks(user_id: str, db: AsyncSession, days: int = 7):
"""清理旧任务记录"""
from sqlalchemy import delete as sql_delete
from datetime import timedelta
cutoff = datetime.now() - timedelta(days=days)
result = await db.execute(
sql_delete(BackgroundTask).where(
BackgroundTask.user_id == user_id,
BackgroundTask.status.in_(["completed", "failed", "cancelled"]),
BackgroundTask.completed_at < cutoff
)
)
if result.rowcount > 0:
await db.commit()
logger.info(f"🧹 清理用户 {user_id[:8]}{result.rowcount} 条旧任务记录")
async def spawn_background_task(
self,
task_id: str,
user_id: str,
task_func: Callable[..., Awaitable],
*args,
**kwargs
):
"""
将任务加入该用户的队列排队执行同一用户FIFO不同用户可并发
Args:
task_id: 任务ID
user_id: 用户ID
task_func: 异步任务函数
*args, **kwargs: 传递给task_func的参数
"""
# 确保该用户的队列和工作协程已启动
queue = self._ensure_user_queue(user_id)
await self._start_user_worker(user_id)
# 将任务放入该用户的队列
await queue.put({
"task_id": task_id,
"task_func": task_func,
"args": {"user_id": user_id, "extra_args": args},
"kwargs": kwargs,
})
queue_size = queue.qsize()
logger.info(f"📥 任务已加入用户 {user_id[:8]} 的队列: {task_id[:8]} (当前队列长度: {queue_size})")
# 更新任务状态,显示排队位置
try:
engine = await get_engine(user_id)
AsyncSessionLocal = async_sessionmaker(
engine, class_=AsyncSession, expire_on_commit=False
)
async with AsyncSessionLocal() as session:
result = await session.execute(
select(BackgroundTask).where(BackgroundTask.id == task_id)
)
task = result.scalar_one_or_none()
if task and task.status == "pending":
if queue_size > 0:
task.status_message = f"排队中,前方还有 {queue_size} 个任务等待..."
else:
task.status_message = "即将开始执行..."
task.progress_details = {"stage": "queued", "queue_size": queue_size}
task.updated_at = datetime.now()
await session.commit()
except Exception as e:
logger.error(f"更新队列位置信息失败: {e}")
def get_queue_size(self, user_id: str = None) -> int:
"""获取队列中等待的任务数量"""
if user_id:
queue = self._user_queues.get(user_id)
return queue.qsize() if queue else 0
# 所有用户队列总数
return sum(q.qsize() for q in self._user_queues.values())
def get_all_queue_info(self) -> Dict[str, int]:
"""获取所有用户的队列信息"""
return {
uid: q.qsize() for uid, q in self._user_queues.items() if q.qsize() > 0
}
# 全局单例
background_task_service = BackgroundTaskService()
+1 -1
View File
@@ -1284,7 +1284,7 @@ class ForeshadowService:
planted_foreshadows = await self.get_planted_foreshadows_for_analysis(db, project_id)
# 每章最多创建的新伏笔数量
MAX_NEW_FORESHADOWS_PER_CHAPTER = 2
MAX_NEW_FORESHADOWS_PER_CHAPTER = 5
new_foreshadow_count = 0
for fs_data in analysis_foreshadows:
+339 -89
View File
@@ -26,15 +26,101 @@ _QUOTE_MAP = {
}
def _is_content_quote(text: str, pos: int) -> bool:
"""
判断字符串值内的 '"' 是否为内容引号需转义而非 JSON 结束引号
合法 JSON 字符串结束引号之后的非空白字符必须是
',' (值分隔) / '}' (关闭对象) / ']' (关闭数组)
如果 '"' 后面不符合这些模式则是 AI 写入的内容引号需要转义
"""
j = pos + 1
# 跳过空格和制表符
while j < len(text) and text[j] in ' \t':
j += 1
if j >= len(text):
return False # 文本末尾,视为结束引号
ch = text[j]
# } 或 ] → 结束引号
if ch in ('}', ']'):
return False
# 换行 → 检查下一行开头判断
if ch == '\n' or ch == '\r':
k = j + (2 if (ch == '\r' and j + 1 < len(text) and text[j + 1] == '\n') else 1)
while k < len(text) and text[k] in ' \t':
k += 1
if k >= len(text):
return False
# 下一行以 " (JSON key) 或 } 或 ] 开头 → 结束引号
if text[k] == '"' or text[k] in ('}', ']'):
return False
return True
# , → 需要检查逗号后面是什么
if ch == ',':
k = j + 1
while k < len(text) and text[k] in ' \t':
k += 1
if k >= len(text):
return False
# 逗号后跟换行 → 检查下一行
if text[k] in ('\n', '\r'):
k2 = k + (2 if (text[k] == '\r' and k + 1 < len(text) and text[k + 1] == '\n') else 1)
while k2 < len(text) and text[k2] in ' \t\n\r':
k2 += 1
if k2 >= len(text):
return False
if text[k2] == '"' or text[k2] in ('}', ']'):
return False
return True
after_comma = text[k]
# 结构性逗号后应为 JSON 值的开头
if after_comma == '"':
return False # 字符串值或 key
if after_comma.isdigit() or after_comma == '-':
return False # 数字
if after_comma in ('{', '['):
return False # 对象/数组
if text[k:k+4] in ('true', 'null'):
return False
if text[k:k+5] == 'false':
return False
# 逗号后不是 JSON 值开头 → 内容逗号,引号是内容引号
return True
# : → 通常在字符串结束后不可能出现,保守处理为结束引号
if ch == ':':
return False
# 其他字符(中文、字母等)→ 内容引号
return True
def _fix_json_string_values(text: str) -> str:
"""
修复JSON字符串值中的常见问题
1. 裸换行符/制表符 转义
2. 字符串值内的中文引号 转义为ASCII引号避免破坏JSON结构
3. 结构位置的中文引号 直接替换为ASCII引号
上下文感知的 JSON 修复区分字符串内外分别处理
AI生成的JSON常在字符串值中插入未转义的换行符和中文引号
此函数遍历文本区分字符串内外分别处理
字符串值内
1. 裸换行符/制表符 转义
2. 中文引号"" 转义为 \\"
3. 未转义的 ASCII 双引号 智能检测内容引号转义结束引号保留
4. 中文逗号/冒号 保留原样是内容字符
结构位置字符串外
1. 中文引号 ASCII 引号
2. 中文逗号 ASCII 逗号
3. 中文冒号 ASCII 冒号
"""
if not text or '"' not in text:
return text
@@ -47,111 +133,234 @@ def _fix_json_string_values(text: str) -> str:
while i < len(text):
c = text[i]
if c == '"' and not in_string:
# 进入字符串
in_string = True
# === 非字符串内(结构位置)===
if not in_string:
# 结构位置的中文标点 → ASCII
if c == '\uff0c': # ,→ ,
result.append(',')
fixed_count += 1
i += 1
continue
if c == '\uff1a': # :→ :
result.append(':')
fixed_count += 1
i += 1
continue
if c in _QUOTE_MAP:
result.append(_QUOTE_MAP[c])
fixed_count += 1
i += 1
continue
# ASCII 双引号 → 进入字符串
if c == '"':
in_string = True
result.append(c)
i += 1
continue
result.append(c)
i += 1
continue
if in_string:
if c == '\\':
# 转义字符,检查下一个字符是否合法
if i + 1 < len(text):
next_c = text[i + 1]
# JSON 合法转义:\" \\ \/ \b \f \n \r \t \uXXXX
if next_c in ('"', '\\', '/', 'b', 'f', 'n', 'r', 't'):
# 合法转义,直接保留
result.append(c)
result.append(next_c)
i += 2
# === 字符串值内 ===
# 转义字符处理
if c == '\\':
if i + 1 < len(text):
next_c = text[i + 1]
if next_c in ('"', '\\', '/', 'b', 'f', 'n', 'r', 't'):
result.append(c)
result.append(next_c)
i += 2
continue
elif next_c == 'u':
if i + 5 < len(text) and all(text[i+2+k] in '0123456789abcdefABCDEF' for k in range(4)):
result.append(text[i:i+6])
i += 6
continue
elif next_c == 'u':
# Unicode 转义 \uXXXX,检查是否有4个十六进制字符
if i + 5 < len(text) and all(text[i+2+k] in '0123456789abcdefABCDEF' for k in range(4)):
result.append(text[i:i+6])
i += 6
continue
else:
# 不完整的unicode转义,去掉反斜杠
result.append(next_c)
fixed_count += 1
i += 2
continue
else:
# 非法转义字符(如 \c \p \d 等),去掉反斜杠只保留字符
result.append(next_c)
fixed_count += 1
i += 2
continue
else:
# 末尾孤立的反斜杠,去掉
result.append(next_c)
fixed_count += 1
i += 1
i += 2
continue
if c == '"':
# 字符串结束
else:
fixed_count += 1
i += 1
continue
# ASCII 双引号 → 智能判断是结束引号还是内容引号
if c == '"':
if _is_content_quote(text, i):
# 内容引号,需要转义
result.append('\\')
result.append('"')
fixed_count += 1
i += 1
continue
else:
# 结束引号
in_string = False
result.append(c)
i += 1
continue
if c == '\n':
# 裸换行符 → 替换为转义换行
result.append('\\')
result.append('n')
fixed_count += 1
i += 1
continue
if c == '\r':
# 裸回车符 → 忽略或替换
if i + 1 < len(text) and text[i + 1] == '\n':
result.append('\\')
result.append('n')
fixed_count += 1
i += 2
else:
result.append('\\')
result.append('n')
fixed_count += 1
i += 1
continue
if c == '\t':
# 裸制表符 → 替换为转义制表符
result.append('\\')
result.append('t')
fixed_count += 1
i += 1
continue
# 字符串值内的中文引号 → 转义为 \"(避免破坏JSON结构)
if c in _QUOTE_MAP:
result.append('\\')
result.append(_QUOTE_MAP[c])
fixed_count += 1
i += 1
continue
# 非字符串内的字符
# 结构位置的中文引号 → 直接替换
if not in_string and c in _QUOTE_MAP:
result.append(_QUOTE_MAP[c])
# 裸换行符 → 转义
if c == '\n':
result.append('\\')
result.append('n')
fixed_count += 1
i += 1
continue
if c == '\r':
if i + 1 < len(text) and text[i + 1] == '\n':
result.append('\\')
result.append('n')
fixed_count += 1
i += 2
else:
result.append('\\')
result.append('n')
fixed_count += 1
i += 1
continue
if c == '\t':
result.append('\\')
result.append('t')
fixed_count += 1
i += 1
continue
# 中文引号处理
if c in _QUOTE_MAP:
mapped = _QUOTE_MAP[c]
if mapped == '"':
# 中文双引号在字符串内需要转义
result.append('\\')
result.append('"')
else:
# 中文单引号在双引号字符串内不需要转义,直接替换
result.append(mapped)
fixed_count += 1
i += 1
continue
# 其他字符(包括中文逗号、中文冒号)→ 保留原样
result.append(c)
i += 1
if fixed_count > 0:
logger.debug(f"✅ 修复了{fixed_count}个JSON问题(控制字符/中文引号")
logger.debug(f"✅ 修复了{fixed_count}个JSON问题(引号/控制字符/中文标点")
return ''.join(result)
def _fix_all_invalid_escapes(text: str) -> str:
"""
兜底修复扫描整个文本中的无效JSON转义序列
_fix_json_string_values 因字符串边界追踪错误而遗漏某些无效转义时
此函数作为兜底不依赖字符串状态追踪扫描整个文本修复所有无效转义
有效JSON转义\\" \\\\ \\/ \\b \\f \\n \\r \\t \\uXXXX
其他 \\X 均为无效转义修复方式为去掉反斜杠只保留字符
"""
if '\\' not in text:
return text
result = []
i = 0
fixed = 0
while i < len(text):
if text[i] == '\\' and i + 1 < len(text):
next_c = text[i + 1]
if next_c in ('"', '\\', '/', 'b', 'f', 'n', 'r', 't'):
# 有效转义,保留
result.append(text[i])
result.append(next_c)
i += 2
continue
elif next_c == 'u':
# Unicode 转义,检查是否有4个十六进制字符
if i + 5 < len(text) and all(
text[i + 2 + k] in '0123456789abcdefABCDEF'
for k in range(4)
):
result.append(text[i:i + 6])
i += 6
continue
else:
# 不完整的unicode转义,去掉反斜杠
result.append(next_c)
fixed += 1
i += 2
continue
else:
# 无效转义(如 \引 \影 \某种 等),去掉反斜杠只保留字符
result.append(next_c)
fixed += 1
i += 2
continue
else:
result.append(text[i])
i += 1
if fixed > 0:
logger.info(f"✅ 兜底修复了{fixed}个无效JSON转义序列")
return ''.join(result)
def _fix_multiple_objects_as_value(text: str) -> str:
"""
修复AI生成的JSON中多个对象作为属性值但未合并的问题
示例
"key": {"a": "1"}, {"b": "2"} "key": {"a": "1", "b": "2"}
AI有时在输出对象类型的属性值时输出了多个独立的对象而不是合并为一个
例如 relationship_changes 字段输出多个角色关系变化时可能出现此问题
此函数检测并合并这些对象
"""
if '{' not in text or '}' not in text:
return text
# 匹配嵌套层级不超过2的对象: { ... } 其中 ... 不含 { 或仅含单层嵌套
nested_obj = r'\{(?:[^{}]|\{(?:[^{}]|\{[^{}]*\})*\})*\}'
# 模式:属性冒号后跟一个对象,然后逗号和另一个对象(没有属性名)
# 即 "key": {obj1}, {obj2} → "key": {obj1, obj2}
pattern = r'(":)\s*(' + nested_obj + r')\s*,\s*(' + nested_obj + r')'
def merge_objects(match):
colon = match.group(1)
obj1_content = match.group(2)[1:-1] # 去掉外层的 { }
obj2_content = match.group(3)[1:-1] # 去掉外层的 { }
# 合并为一个对象
return f'{colon} {{{obj1_content}, {obj2_content}}}'
prev = None
count = 0
max_iterations = 10
while prev != text and count < max_iterations:
prev = text
text = re.sub(pattern, merge_objects, text)
count += 1
if count > 1:
logger.info(f"✅ 修复了{count - 1}处多对象属性值合并")
return text
def clean_json_response(text: str) -> str:
"""清洗 AI 返回的 JSON(改进版 - 流式安全)"""
try:
@@ -162,11 +371,8 @@ def clean_json_response(text: str) -> str:
original_length = len(text)
logger.debug(f"🔍 开始清洗JSON,原始长度: {original_length}")
# 替换中文逗号/冒号(AI可能在JSON结构位置使用,全局替换是安全的)
text = text.replace('\uff0c', ',') # ,→ ,
text = text.replace('\uff1a', ':') # :→ :
# 修复JSON中的中文引号和裸控制字符(上下文感知,区分字符串内外)
# 上下文感知修复:中文引号/逗号/冒号、裸控制字符、未转义的内容引号
# (区分字符串内外:结构位置替换为ASCII,字符串内保留或转义)
text = _fix_json_string_values(text)
# 去除 markdown 代码块
@@ -286,9 +492,35 @@ def clean_json_response(text: str) -> str:
json.loads(result)
logger.debug(f"✅ 清洗后JSON验证成功")
except json.JSONDecodeError as e:
logger.error(f" 清洗后JSON仍然无效: {e}")
logger.debug(f" 结果预览: {result[:500]}")
logger.debug(f" 结果结尾: ...{result[-200:]}")
logger.warning(f"⚠️ 清洗后JSON仍然无效: {e},尝试修复结构性问题...")
# 修复1:合并多对象属性值(AI可能输出 "key": {a:1}, {b:2}
result = _fix_multiple_objects_as_value(result)
try:
json.loads(result)
logger.info(f"✅ 修复多对象属性值后JSON验证成功")
except json.JSONDecodeError:
pass # 继续尝试其他修复
else:
return result
# 修复2:兜底修复无效转义序列(不依赖字符串边界追踪)
logger.warning(f"⚠️ 继续尝试兜底修复无效转义...")
result = _fix_all_invalid_escapes(result)
try:
json.loads(result)
logger.info(f"✅ 兜底修复后JSON验证成功")
except json.JSONDecodeError as e2:
# 修复3:再次尝试合并多对象属性值(转义修复后可能产生新的合并机会)
result = _fix_multiple_objects_as_value(result)
try:
json.loads(result)
logger.info(f"✅ 二次修复后JSON验证成功")
except json.JSONDecodeError as e3:
logger.error(f"❌ 所有修复后JSON仍然无效: {e3}")
logger.debug(f" 结果预览: {result[:500]}")
logger.debug(f" 结果结尾: ...{result[-200:]}")
return result
@@ -339,6 +571,16 @@ def loads_json(text: str) -> Any:
except (json.JSONDecodeError, Exception):
pass
# 兜底修复无效转义序列后重试
fixed_text = _fix_all_invalid_escapes(text)
if fixed_text != text:
try:
result = json.loads(fixed_text)
logger.info("✅ 兜底修复无效转义后json.loads成功")
return result
except (json.JSONDecodeError, Exception):
pass
# json5 容错解析
if HAS_JSON5:
try:
@@ -347,6 +589,14 @@ def loads_json(text: str) -> Any:
logger.info("✅ json5容错解析成功")
return result
except Exception as e5:
# json5也失败,尝试对修复后的文本使用json5
if fixed_text != text:
try:
result = json5.loads(fixed_text)
logger.info("✅ 兜底修复无效转义后json5容错解析成功")
return result
except Exception:
pass
logger.error(f"❌ json5容错解析也失败: {e5}")
# 最终失败,抛出标准异常