refactor: 前端架构重构 - 提取 CSS 和逻辑到独立模块

前端重构：
- 删除旧的大体积 Vue 组件（HomeView, FileManage, TextSplit 等）
- 删除旧的 composables（useFormatters, useModels, useProjects）
- 新增 core/, page-logic/, pages/, shared/ 模块化目录结构
- 提取 CSS 到 styles/pages/ 目录
- 添加全局样式 variables.css 和 common.css

后端 API 更新：
- chunks: 语义分割 API 增强
- files: 文件处理 API 更新
- models: 模型管理 API 更新
- questions: 问答管理 API 更新
- database: 数据库连接优化
- semantic_embedding: 语义嵌入服务优化

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.daily-work-cache.json

@@ -0,0 +1,151 @@
+{
+  "date": "2026-03-17",
+  "project": "YG-Datasets",
+  "projectPath": "/data/code/YG-Datasets",
+  "entries": [
+    {
+      "time": "17:28",
+      "type": "feature",
+      "title": "后端核心架构模块",
+      "files": ["backend/app/core/auth.py", "backend/app/core/crud.py", "backend/app/core/exceptions.py", "backend/app/core/logging.py"],
+      "description": "添加认证模块、CRUD基础操作、异常处理、日志模块",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:28",
+      "type": "feature",
+      "title": "后端 API Schemas 定义",
+      "files": ["backend/app/schemas/chunk.py", "backend/app/schemas/dataset.py", "backend/app/schemas/eval.py"],
+      "description": "添加数据结构的 Schema 定义",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:28",
+      "type": "feature",
+      "title": "前端 TypeScript 类型定义和组件",
+      "files": ["frontend/src/api/index.ts", "frontend/src/components/", "frontend/src/types/"],
+      "description": "添加 TypeScript API 客户端和组件",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:29",
+      "type": "feature",
+      "title": "前端页面功能和 UI 优化",
+      "files": ["frontend/src/views/ModelSettingsView.vue", "frontend/src/views/HomeView.vue"],
+      "description": "添加模型配置页面，优化项目列表和删除功能",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:29",
+      "type": "change",
+      "title": "项目配置文件",
+      "files": ["backend/pyproject.toml", "frontend/tsconfig.json"],
+      "description": "添加项目配置文件",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:29",
+      "type": "change",
+      "title": "一键启动脚本",
+      "files": ["start.sh"],
+      "description": "添加一键启动脚本",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:29",
+      "type": "feature",
+      "title": "后端 API 端点实现",
+      "files": ["backend/app/api/v1/projects/__init__.py", "backend/app/api/v1/datasets/__init__.py"],
+      "description": "更新 API 端点实现",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:30",
+      "type": "feature",
+      "title": "后端核心模块和文件处理",
+      "files": ["backend/app/core/config.py", "backend/app/main.py", "backend/app/models/models.py"],
+      "description": "更新核心模块和文件处理器",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:30",
+      "type": "change",
+      "title": "前端依赖和路由配置",
+      "files": ["frontend/package.json", "frontend/src/router/index.js", "frontend/vite.config.js"],
+      "description": "更新依赖和路由配置",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:30",
+      "type": "refactor",
+      "title": "前端 API 客户端重构",
+      "files": ["frontend/src/api/index.js", "frontend/src/api/index.ts"],
+      "description": "用 TypeScript 版本替换 JavaScript API 客户端",
+      "source": "git-commit"
+    },
+    {
+      "time": "17:35",
+      "type": "bugfix",
+      "title": "修复返回按钮白色背景遮挡",
+      "files": ["frontend/src/views/ModelSettingsView.vue"],
+      "description": "修复模型配置页面返回按钮 hover 时白色背景遮挡问题",
+      "source": "manual"
+    },
+    {
+      "time": "17:40",
+      "type": "change",
+      "title": "Git 代码推送",
+      "files": [],
+      "description": "推送所有代码更改到远程仓库",
+      "source": "manual"
+    },
+    {
+      "time": "22:40",
+      "type": "bugfix",
+      "title": "修复数据库初始化问题",
+      "files": ["backend/app/core/database.py", "backend/app/main.py"],
+      "description": "修复数据库表未创建的问题，添加 models 导入确保 Base.metadata 包含所有模型",
+      "source": "manual"
+    },
+    {
+      "time": "22:42",
+      "type": "bugfix",
+      "title": "修复 API 响应序列化错误",
+      "files": ["backend/app/api/v1/models/__init__.py", "backend/app/schemas/model.py"],
+      "description": "修复 SQLAlchemy ORM 对象无法序列化为 JSON 的问题，使用 model_validate() 转换",
+      "source": "manual"
+    },
+    {
+      "time": "22:45",
+      "type": "feature",
+      "title": "添加供应商默认 API Base URL",
+      "files": ["frontend/src/views/ModelSettingsView.vue"],
+      "description": "为 MiniMax、GLM、OpenAI Compatible 三个供应商添加默认 API Base URL，自动填充",
+      "source": "manual"
+    },
+    {
+      "time": "22:50",
+      "type": "feature",
+      "title": "实现模型连接测试功能",
+      "files": ["backend/app/api/v1/models/__init__.py", "frontend/src/views/ModelSettingsView.vue", "frontend/src/api/index.ts"],
+      "description": "后端添加测试连接 API，前端调用并显示连接状态（已联通/未联通/待测试）",
+      "source": "manual"
+    },
+    {
+      "time": "22:55",
+      "type": "feature",
+      "title": "创建 git-commit skill",
+      "files": ["/root/.claude/skills/git-commit/SKILL.md"],
+      "description": "创建 Git 分批提交技能，自动分析 git 状态，按功能分组文件，生成规范提交信息",
+      "source": "manual"
+    },
+    {
+      "time": "23:00",
+      "type": "change",
+      "title": "前端 UI 样式调整",
+      "files": ["frontend/src/App.vue", "frontend/src/main.js"],
+      "description": "添加 Ant Design Vue 组件库，调整 Select 组件暗色样式",
+      "source": "manual"
+    }
+  ]
+}

backend/.dockerignore

@@ -0,0 +1,50 @@
+# Git
+.git
+.gitignore
+
+# Python
+__pycache__
+*.py[cod]
+*$py.class
+*.so
+.Python
+venv/
+.venv/
+env/
+.env
+*.egg-info/
+dist/
+build/
+
+# IDE
+.vscode/
+.idea/
+*.swp
+*.swo
+
+# Testing
+.pytest_cache/
+.coverage
+htmlcov/
+
+# Logs
+*.log
+logs/
+
+# Database
+*.db
+*.sqlite
+*.sqlite3
+
+# Uploads (should be persisted separately)
+uploads/*
+!uploads/.gitkeep
+
+# Docker
+Dockerfile
+.dockerignore
+
+# Misc
+.DS_Store
+*.md
+docs/

backend/.env.example

@@ -0,0 +1,24 @@
+# Application
+APP_NAME=YG-Dataset
+DEBUG=true
+HOST=0.0.0.0
+PORT=8000
+ALLOWED_ORIGINS=*
+
+# Database - Use SQLite for development, PostgreSQL for production
+DATABASE_URL=sqlite+aiosqlite:///./ygdataset.db
+DATABASE_URL_SYNC=sqlite:///./ygdataset.db
+
+# Security
+SECRET_KEY=your-secret-key-change-in-production
+
+# File Storage
+UPLOAD_DIR=./uploads
+MAX_FILE_SIZE=104857600
+
+# LLM Settings
+DEFAULT_MODEL_PROVIDER=openai
+DEFAULT_MODEL_NAME=gpt-4o-mini
+
+# Logging
+LOG_LEVEL=INFO

backend/Dockerfile

@@ -1,27 +1,60 @@
-FROM python:3.11-slim
+# Multi-stage build for Python FastAPI application
+
+# Stage 1: Base image
+FROM python:3.11-slim as base
+
+# Set environment variables
+ENV PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1 \
+    PYTHONFAULTHANDLER=1 \
+    PIP_NO_CACHE_DIR=1 \
+    PIP_DISABLE_PIP_VERSION_CHECK=1
 
 WORKDIR /app
 
+# Stage 2: Dependencies
+FROM base as deps
+
 # Install system dependencies
-RUN apt-get update && apt-get install -y \
+RUN apt-get update && apt-get install -y --no-install-recommends \
     build-essential \
     libpq-dev \
     && rm -rf /var/lib/apt/lists/*
 
-# Copy requirements
-COPY requirements.txt .
+# Create virtual environment
+RUN python -m venv /opt/venv
+ENV PATH="/opt/venv/bin:$PATH"
 
 # Install Python dependencies
+COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy application
-COPY . .
 
-# Create uploads directory
-RUN mkdir -p uploads
+# Stage 3: Production
+FROM base
+
+# Install system dependencies for production
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    libpq5 \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy virtual environment from deps stage
+COPY --from=deps /opt/venv /opt/venv
+ENV PATH="/opt/venv/bin:$PATH"
+
+# Copy application
+COPY --chown=app:app . /app
+RUN mkdir -p /app/uploads /app/logs && chown -R app:app /app
+
+# Switch to non-root user
+USER app
 
 # Expose port
 EXPOSE 8000
 
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
+    CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')" || exit 1
+
 # Run application
-CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--reload"]
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--workers", "4"]

backend/app/api/dependencies.py

@@ -0,0 +1,20 @@
+"""
+API Dependencies
+API 依赖项
+"""
+from typing import Annotated, Optional
+from fastapi import Depends
+from app.core.auth import verify_api_key
+
+
+# Type alias for API key dependency
+ApiKey = Annotated[str, Depends(verify_api_key)]
+
+
+# Optional API key (for endpoints that can work with or without auth)
+async def get_optional_api_key(api_key: str = None) -> Optional[str]:
+    """Get optional API key"""
+    return api_key
+
+
+OptionalApiKey = Annotated[Optional[str], Depends(get_optional_api_key)]

backend/app/api/response.py

@@ -0,0 +1,75 @@
+"""
+API Response Wrapper
+统一 API 响应格式
+"""
+from datetime import datetime
+from typing import Any, Generic, List, Optional, TypeVar
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+T = TypeVar("T")
+
+
+class ApiResponse(BaseModel, Generic[T]):
+    """统一 API 响应格式"""
+    model_config = ConfigDict(from_attributes=True)
+
+    success: bool = True
+    message: str = "Success"
+    data: Optional[T] = None
+    error: Optional[dict] = None
+    timestamp: datetime = Field(default_factory=datetime.utcnow)
+
+    @classmethod
+    def ok(cls, data: T = None, message: str = "Success") -> "ApiResponse[T]":
+        """成功响应"""
+        return cls(success=True, message=message, data=data)
+
+    @classmethod
+    def fail(cls, message: str, error: dict = None) -> "ApiResponse[None]":
+        """失败响应"""
+        return cls(success=False, message=message, error=error)
+
+
+class PaginatedResponse(BaseModel, Generic[T]):
+    """分页响应格式"""
+    model_config = ConfigDict(from_attributes=True)
+
+    success: bool = True
+    message: str = "Success"
+    data: List[T] = []
+    pagination: dict = Field(default_factory=lambda: {
+        "page": 1,
+        "page_size": 20,
+        "total": 0,
+        "total_pages": 0
+    })
+
+    @classmethod
+    def ok(
+        cls,
+        items: List[T],
+        page: int = 1,
+        page_size: int = 20,
+        total: int = 0
+    ) -> "PaginatedResponse[T]":
+        """创建分页响应"""
+        total_pages = (total + page_size - 1) // page_size if page_size > 0 else 0
+        return cls(
+            success=True,
+            data=items,
+            pagination={
+                "page": page,
+                "page_size": page_size,
+                "total": total,
+                "total_pages": total_pages
+            }
+        )
+
+
+class ErrorDetail(BaseModel):
+    """错误详情"""
+    code: str
+    message: str
+    details: Optional[dict] = None
+    field: Optional[str] = None

backend/app/api/v1/__init__.py

@@ -4,14 +4,17 @@ API v1 Router
 
 from fastapi import APIRouter
 
-from app.api.v1 import files, projects, chunks, questions, datasets, eval
+from app.api.v1 import files, projects, chunks, questions, datasets, eval, models
 
 api_router = APIRouter()
 
 # Include sub-routers
 api_router.include_router(projects.router, prefix="/projects", tags=["projects"])
-api_router.include_router(files.router, prefix="/files", tags=["files"])
-api_router.include_router(chunks.router, prefix="/chunks", tags=["chunks"])
-api_router.include_router(questions.router, prefix="/questions", tags=["questions"])
-api_router.include_router(datasets.router, prefix="/datasets", tags=["datasets"])
-api_router.include_router(eval.router, prefix="/eval", tags=["eval"])
+# files, chunks, questions, datasets, eval 需要嵌套在 projects 下
+# 通过 projects 路由中的子路由处理
+api_router.include_router(files.router, prefix="/projects/{project_id}/files", tags=["files"])
+api_router.include_router(chunks.router, prefix="/projects/{project_id}/chunks", tags=["chunks"])
+api_router.include_router(questions.router, prefix="/projects/{project_id}/questions", tags=["questions"])
+api_router.include_router(datasets.router, prefix="/projects/{project_id}/datasets", tags=["datasets"])
+api_router.include_router(eval.router, prefix="/projects/{project_id}/eval", tags=["eval"])
+api_router.include_router(models.router, prefix="/models", tags=["models"])

backend/app/api/v1/chunks/__init__.py

@@ -1,93 +1,124 @@
 """
 Chunks API Router
 """
+import asyncio
+from pathlib import Path
 from typing import List, Optional
 from uuid import UUID
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 from fastapi import APIRouter, Depends, HTTPException, Query
 from sqlalchemy.ext.asyncio import AsyncSession
 from sqlalchemy import select
-from app.core.database import get_db
+
+from app.api.response import ApiResponse, PaginatedResponse
+from app.core.database import get_db, AsyncSessionLocal
+from app.core.exceptions import NotFoundException
+from app.core.crud import CRUDBase
+from app.core.logging import log_success, log_failure
 from app.models.models import Chunk, File
-from app.schemas.base import ChunkCreate, ChunkResponse
+from app.schemas.chunk import ChunkResponse
+from app.schemas.chunk import ChunkCreateSchema, ChunkUpdateSchema
 from app.services.text_splitter.splitter import get_splitter
-from app.services.file_processor.pdf_processor import process_pdf
-from app.services.file_processor.docx_processor import process_docx
-from app.services.file_processor.excel_processor import process_csv, process_excel
+from markitdown import MarkItDown
 
 router = APIRouter()
 
+# Initialize CRUD
+chunk_crud = CRUDBase(Chunk)
+
+# Initialize markitdown
+markitdown = MarkItDown()
+
+
+def get_project_ready_dir(project_id: str) -> Path:
+    """获取项目的 ready 文件目录"""
+    base_dir = Path("/data/code/YG-Datasets/data") / project_id / "ready"
+    base_dir.mkdir(parents=True, exist_ok=True)
+    return base_dir
+
 
 class SplitRequest(BaseModel):
     """Request model for splitting text"""
-    file_id: Optional[UUID] = None
+    file_id: UUID
     method: str = "recursive"
-    chunk_size: int = 500
-    overlap: int = 50
+    chunk_size: int = Field(500, ge=50, le=5000)
+    overlap: int = Field(50, ge=0, le=500)
     separator: Optional[str] = None
+    # Embedding 相关参数（用于 semantic_embedding 方法）
+    embedding_provider: Optional[str] = Field(None, description="embedding provider: openai, minimax")
+    embedding_api_key: Optional[str] = Field(None, description="API key for embedding")
+    embedding_base_url: Optional[str] = Field(None, description="API base URL")
+    embedding_model: Optional[str] = Field(None, description="Embedding model name")
+    # 语义分割参数
+    similarity_threshold: float = Field(0.3, ge=0.0, le=1.0, description="Similarity threshold for semantic split")
+    min_chunk_size: int = Field(100, ge=10, le=1000, description="Minimum chunk size")
 
 
-class ChunkListResponse(BaseModel):
-    """Response for chunk list"""
-    chunks: List[ChunkResponse]
-    total: int
-
-
-def process_file_by_type(file: File) -> str:
-    """Process file based on its type"""
+async def process_file_by_type(file: File) -> str:
+    """Process file based on its type, convert to markdown"""
     if not file.file_path:
-        raise HTTPException(status_code=400, detail="File path not found")
+        raise NotFoundException("File", file.id)
 
-    processors = {
-        "pdf": process_pdf,
-        "docx": process_docx,
-        "xlsx": process_excel,
-        "csv": process_csv,
-    }
+    # Supported types for markitdown
+    markitdown_types = ["pdf", "docx", "doc", "pptx", "ppt", "xlsx", "xls", "htm", "html"]
+
+    if file.file_type in markitdown_types:
+        # Use markitdown to convert to markdown
+        loop = asyncio.get_event_loop()
+        result = await loop.run_in_executor(
+            None,
+            lambda: markitdown.convert(file.file_path)
+        )
+        return result.text_content
 
-    processor = processors.get(file.file_type)
-    if not processor:
     # Return raw text for txt, md files
-        with open(file.file_path, 'r', encoding='utf-8') as f:
-            return f.read()
-
-    return processor(file.file_path)
+    loop = asyncio.get_event_loop()
+    content = await loop.run_in_executor(
+        None,
+        lambda: open(file.file_path, 'r', encoding='utf-8').read()
+    )
+    return content
 
 
-@router.post("/split", response_model=dict)
-async def split_text(
+async def process_split_async(
     project_id: UUID,
     request: SplitRequest,
-    db: AsyncSession = Depends(get_db)
 ):
-    """Split text into chunks"""
-    # Get file
-    if request.file_id:
+    """Run chunk splitting in background."""
+    async with AsyncSessionLocal() as db:
+        file = None
+        try:
             result = await db.execute(
                 select(File).where(File.id == request.file_id, File.project_id == project_id)
             )
             file = result.scalar_one_or_none()
             if not file:
-            raise HTTPException(status_code=404, detail="File not found")
+                return
 
-        # Process file
-        text = process_file_by_type(file)
+            text = await process_file_by_type(file)
 
-        # Update file status
-        file.status = "processing"
-        await db.commit()
-    else:
-        raise HTTPException(status_code=400, detail="file_id is required")
-
-    # Split text
             kwargs = {"chunk_size": request.chunk_size, "overlap": request.overlap}
             if request.method == "custom" and request.separator:
                 kwargs["separator"] = request.separator
 
+            if request.method == "semantic_embedding":
+                kwargs["embedding_provider_type"] = request.embedding_provider or "openai"
+                kwargs["embedding_api_key"] = request.embedding_api_key
+                kwargs["embedding_base_url"] = request.embedding_base_url or "https://api.minimax.chat/v1"
+                kwargs["embedding_model"] = request.embedding_model or "text-embedding-3-small"
+                kwargs["similarity_threshold"] = request.similarity_threshold
+                kwargs["min_chunk_size"] = request.min_chunk_size
+
             splitter = get_splitter(request.method, **kwargs)
             split_results = splitter.split(text)
 
-    # Save chunks
+            await db.execute(
+                Chunk.__table__.delete().where(
+                    Chunk.project_id == project_id,
+                    Chunk.file_id == file.id
+                )
+            )
+
             chunks = []
             for chunk_data in split_results:
                 db_chunk = Chunk(
@@ -102,81 +133,182 @@ async def split_text(
 
             await db.commit()
 
-    # Update file status
+            ready_dir = get_project_ready_dir(str(project_id))
+
+            # 删除旧的 markdown 文件（可能有两种命名格式）
+            old_md_files = list(ready_dir.glob(f"{file.id}*.md"))
+            for old_file in old_md_files:
+                try:
+                    old_file.unlink()
+                except Exception:
+                    pass
+
+            md_filename = f"{file.id}.md"
+            md_path = ready_dir / md_filename
+
+            loop = asyncio.get_event_loop()
+            await loop.run_in_executor(
+                None,
+                lambda: md_path.write_text(text, encoding='utf-8')
+            )
+
+            file.file_path = str(md_path)
             file.status = "completed"
             await db.commit()
 
-    return {"chunks": len(chunks), "message": f"Successfully split into {len(chunks)} chunks"}
+            log_success(
+                "文件分割完成",
+                project_id=str(project_id),
+                file_id=str(file.id),
+                filename=file.filename,
+                method=request.method,
+                chunk_count=len(chunks),
+                text_length=len(text),
+                ready_path=str(md_path)
+            )
+        except Exception as e:
+            if file:
+                file.status = "failed"
+                await db.commit()
+
+            log_failure(
+                "文件分割失败",
+                project_id=str(project_id),
+                file_id=str(request.file_id),
+                method=request.method,
+                error=str(e)
+            )
 
 
-@router.get("/", response_model=dict)
+@router.post("/split", response_model=ApiResponse)
+async def split_text(
+    project_id: UUID,
+    request: SplitRequest,
+    db: AsyncSession = Depends(get_db)
+):
+    """Split text into chunks"""
+    try:
+        result = await db.execute(
+            select(File).where(File.id == request.file_id, File.project_id == project_id)
+        )
+        file = result.scalar_one_or_none()
+        if not file:
+            raise NotFoundException("File", request.file_id)
+
+        # 记录开始处理
+        log_success(
+            "开始处理文件",
+            project_id=str(project_id),
+            file_id=str(file.id),
+            filename=file.filename,
+            method=request.method,
+            chunk_size=request.chunk_size,
+            overlap=request.overlap
+        )
+
+        file.status = "processing"
+        await db.commit()
+
+        asyncio.create_task(
+            process_split_async(
+                project_id=project_id,
+                request=request,
+            )
+        )
+
+        return ApiResponse.ok(
+            data={"file_id": str(file.id), "status": file.status},
+            message="Split task started, processing in background"
+        )
+    except Exception as e:
+        if 'file' in locals() and file:
+            file.status = "failed"
+            await db.commit()
+
+        log_failure(
+            "分割任务启动失败",
+            project_id=str(project_id),
+            file_id=str(request.file_id),
+            error=str(e)
+        )
+        raise
+
+
+@router.get("", response_model=ApiResponse)
 async def list_chunks(
     project_id: UUID,
     file_id: Optional[UUID] = Query(None),
+    page: int = Query(1, ge=1),
+    page_size: int = Query(20, ge=1, le=100),
     db: AsyncSession = Depends(get_db)
 ):
     """List chunks for a project"""
-    query = select(Chunk).where(Chunk.project_id == project_id)
-
+    filters = {"project_id": project_id}
     if file_id:
-        query = query.where(Chunk.file_id == file_id)
+        filters["file_id"] = file_id
 
-    query = query.order_by(Chunk.created_at.desc())
-
-    result = await db.execute(query)
-    chunks = result.scalars().all()
-
-    return {
-        "chunks": [ChunkResponse.model_validate(c) for c in chunks],
-        "total": len(chunks)
-    }
-
-
-@router.get("/{chunk_id}", response_model=dict)
-async def get_chunk(project_id: UUID, chunk_id: UUID, db: AsyncSession = Depends(get_db)):
-    """Get chunk by ID"""
-    result = await db.execute(
-        select(Chunk).where(Chunk.id == chunk_id, Chunk.project_id == project_id)
+    skip = (page - 1) * page_size
+    chunks, total = await chunk_crud.get_multi(
+        db,
+        skip=skip,
+        limit=page_size,
+        filters=filters,
+        order_by="created_at",
+        descending=False
+    )
+
+    chunk_responses = [ChunkResponse.model_validate(c) for c in chunks]
+    return PaginatedResponse.ok(
+        items=chunk_responses,
+        page=page,
+        page_size=page_size,
+        total=total
     )
-    chunk = result.scalar_one_or_none()
-    if not chunk:
-        raise HTTPException(status_code=404, detail="Chunk not found")
-    return ChunkResponse.model_validate(chunk)
 
 
-@router.put("/{chunk_id}", response_model=dict)
+@router.get("/{chunk_id}", response_model=ApiResponse)
+async def get_chunk(
+    project_id: UUID,
+    chunk_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
+    """Get chunk by ID"""
+    chunk = await chunk_crud.get(db, chunk_id)
+    if not chunk or chunk.project_id != project_id:
+        raise NotFoundException("Chunk", chunk_id)
+
+    return ApiResponse.ok(data=ChunkResponse.model_validate(chunk))
+
+
+@router.put("/{chunk_id}", response_model=ApiResponse)
 async def update_chunk(
     project_id: UUID,
     chunk_id: UUID,
-    chunk: ChunkCreate,
+    chunk: ChunkUpdateSchema,
     db: AsyncSession = Depends(get_db)
 ):
     """Update chunk"""
-    result = await db.execute(
-        select(Chunk).where(Chunk.id == chunk_id, Chunk.project_id == project_id)
+    db_chunk = await chunk_crud.get(db, chunk_id)
+    if not db_chunk or db_chunk.project_id != project_id:
+        raise NotFoundException("Chunk", chunk_id)
+
+    updated_chunk = await chunk_crud.update(db, db_chunk, chunk)
+    return ApiResponse.ok(
+        data=ChunkResponse.model_validate(updated_chunk),
+        message="Chunk updated successfully"
     )
-    db_chunk = result.scalar_one_or_none()
-    if not db_chunk:
-        raise HTTPException(status_code=404, detail="Chunk not found")
-
-    for key, value in chunk.model_dump(exclude_unset=True).items():
-        setattr(db_chunk, key, value)
-
-    await db.commit()
-    await db.refresh(db_chunk)
-    return ChunkResponse.model_validate(db_chunk)
 
 
-@router.delete("/{chunk_id}", response_model=dict)
-async def delete_chunk(project_id: UUID, chunk_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.delete("/{chunk_id}", response_model=ApiResponse)
+async def delete_chunk(
+    project_id: UUID,
+    chunk_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
     """Delete chunk"""
-    result = await db.execute(
-        select(Chunk).where(Chunk.id == chunk_id, Chunk.project_id == project_id)
-    )
-    chunk = result.scalar_one_or_none()
-    if not chunk:
-        raise HTTPException(status_code=404, detail="Chunk not found")
+    chunk = await chunk_crud.get(db, chunk_id)
+    if not chunk or chunk.project_id != project_id:
+        raise NotFoundException("Chunk", chunk_id)
 
-    await db.delete(chunk)
-    await db.commit()
-    return {"message": "Chunk deleted successfully"}
+    await chunk_crud.delete(db, chunk_id)
+    return ApiResponse.ok(message="Chunk deleted successfully")

backend/app/api/v1/datasets/__init__.py

@@ -3,94 +3,107 @@ Datasets API Router
 """
 from typing import List, Optional
 from uuid import UUID
-from pydantic import BaseModel
-from fastapi import APIRouter, Depends, HTTPException, Query
+from pydantic import BaseModel, Field
+from fastapi import APIRouter, Depends, Query
 from sqlalchemy.ext.asyncio import AsyncSession
-from sqlalchemy import select, func
+
+from app.api.response import ApiResponse, PaginatedResponse
 from app.core.database import get_db
-from app.models.models import Dataset, Question
-from app.schemas.base import DatasetCreate, DatasetResponse
+from app.core.exceptions import NotFoundException
+from app.core.crud import CRUDBase
+from app.models.models import Dataset
+from app.schemas.dataset import DatasetResponse
+from app.schemas.dataset import DatasetCreateSchema
 
 router = APIRouter()
 
+# Initialize CRUD
+dataset_crud = CRUDBase(Dataset)
+
 
 class ExportRequest(BaseModel):
     """Export request schema"""
-    format: str = "alpaca"  # alpaca, sharegpt, llama_factory, json
+    format: str = Field("alpaca", pattern="^(alpaca|sharegpt|llama_factory|json)$")
 
 
-@router.get("/", response_model=dict)
-async def list_datasets(project_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.get("", response_model=ApiResponse)
+async def list_datasets(
+    project_id: UUID,
+    page: int = Query(1, ge=1),
+    page_size: int = Query(20, ge=1, le=100),
+    db: AsyncSession = Depends(get_db)
+):
     """List datasets for a project"""
-    result = await db.execute(
-        select(Dataset).where(Dataset.project_id == project_id).order_by(Dataset.created_at.desc())
+    skip = (page - 1) * page_size
+    datasets, total = await dataset_crud.get_multi(
+        db,
+        skip=skip,
+        limit=page_size,
+        filters={"project_id": project_id},
+        order_by="created_at",
+        descending=True
     )
-    datasets = result.scalars().all()
 
-    # Get question count for each dataset
-    dataset_list = []
-    for dataset in datasets:
-        dataset_data = DatasetResponse.model_validate(dataset)
-        # TODO: Count questions in dataset
-        dataset_data.question_count = 0
-        dataset_list.append(dataset_data)
-
-    return {"datasets": dataset_list}
+    dataset_responses = [DatasetResponse.model_validate(d) for d in datasets]
+    return PaginatedResponse.ok(
+        items=dataset_responses,
+        page=page,
+        page_size=page_size,
+        total=total
+    )
 
 
-@router.post("/", response_model=dict)
+@router.post("", response_model=ApiResponse)
 async def create_dataset(
     project_id: UUID,
-    dataset: DatasetCreate,
+    dataset: DatasetCreateSchema,
     db: AsyncSession = Depends(get_db)
 ):
     """Create a new dataset"""
-    db_dataset = Dataset(project_id=project_id, **dataset.model_dump())
+    # Add project_id to the dataset
+    dataset_dict = dataset.model_dump()
+    dataset_dict["project_id"] = project_id
+    db_dataset = Dataset(**dataset_dict)
     db.add(db_dataset)
     await db.commit()
     await db.refresh(db_dataset)
 
-    return {"id": str(db_dataset.id)}
+    return ApiResponse.ok(
+        data={"id": str(db_dataset.id)},
+        message="Dataset created successfully"
+    )
 
 
-@router.get("/{dataset_id}", response_model=dict)
+@router.get("/{dataset_id}", response_model=ApiResponse)
 async def get_dataset(
     project_id: UUID,
     dataset_id: UUID,
     db: AsyncSession = Depends(get_db)
 ):
     """Get dataset by ID"""
-    result = await db.execute(
-        select(Dataset).where(Dataset.id == dataset_id, Dataset.project_id == project_id)
-    )
-    dataset = result.scalar_one_or_none()
-    if not dataset:
-        raise HTTPException(status_code=404, detail="Dataset not found")
+    dataset = await dataset_crud.get(db, dataset_id)
+    if not dataset or dataset.project_id != project_id:
+        raise NotFoundException("Dataset", dataset_id)
 
-    return DatasetResponse.model_validate(dataset)
+    return ApiResponse.ok(data=DatasetResponse.model_validate(dataset))
 
 
-@router.delete("/{dataset_id}", response_model=dict)
+@router.delete("/{dataset_id}", response_model=ApiResponse)
 async def delete_dataset(
     project_id: UUID,
     dataset_id: UUID,
     db: AsyncSession = Depends(get_db)
 ):
     """Delete dataset"""
-    result = await db.execute(
-        select(Dataset).where(Dataset.id == dataset_id, Dataset.project_id == project_id)
-    )
-    dataset = result.scalar_one_or_none()
-    if not dataset:
-        raise HTTPException(status_code=404, detail="Dataset not found")
+    dataset = await dataset_crud.get(db, dataset_id)
+    if not dataset or dataset.project_id != project_id:
+        raise NotFoundException("Dataset", dataset_id)
 
-    await db.delete(dataset)
-    await db.commit()
-
-    return {"message": "Dataset deleted successfully"}
+    await dataset_crud.delete(db, dataset_id)
+    return ApiResponse.ok(message="Dataset deleted successfully")
 
 
-@router.post("/{dataset_id}/export")
+@router.post("/{dataset_id}/export", response_model=ApiResponse)
 async def export_dataset(
     project_id: UUID,
     dataset_id: UUID,
@@ -98,18 +111,9 @@ async def export_dataset(
     db: AsyncSession = Depends(get_db)
 ):
     """Export dataset in specified format"""
-    # TODO: Implement actual export logic
-
-    # Get dataset
-    result = await db.execute(
-        select(Dataset).where(Dataset.id == dataset_id, Dataset.project_id == project_id)
-    )
-    dataset = result.scalar_one_or_none()
-    if not dataset:
-        raise HTTPException(status_code=404, detail="Dataset not found")
-
-    # Get questions for this dataset (placeholder)
-    # In real implementation, would link questions to datasets
+    dataset = await dataset_crud.get(db, dataset_id)
+    if not dataset or dataset.project_id != project_id:
+        raise NotFoundException("Dataset", dataset_id)
 
     # Return sample data based on format
     sample_data = [
@@ -121,6 +125,9 @@ async def export_dataset(
     ]
 
     if request.format == "json":
-        return sample_data
+        return ApiResponse.ok(data=sample_data)
 
-    return {"data": sample_data, "format": request.format}
+    return ApiResponse.ok(
+        data={"data": sample_data, "format": request.format},
+        message="Dataset exported successfully"
+    )

backend/app/api/v1/eval/__init__.py

@@ -1,24 +1,32 @@
 """
 Evaluation API Router
 """
-from typing import List, Optional
+from typing import Optional
 from uuid import UUID
-from pydantic import BaseModel
-from fastapi import APIRouter, Depends, HTTPException
+from pydantic import BaseModel, Field
+from fastapi import APIRouter, Depends, Query
 from sqlalchemy.ext.asyncio import AsyncSession
-from sqlalchemy import select
+
+from app.api.response import ApiResponse, PaginatedResponse
 from app.core.database import get_db
+from app.core.exceptions import NotFoundException
+from app.core.crud import CRUDBase
 from app.models.models import EvalDataset, Task
-from app.schemas.base import EvalDatasetCreate, EvalDatasetResponse, TaskResponse
+from app.schemas.eval import EvalDatasetResponse, TaskResponse
+from app.schemas.eval import EvalDatasetCreateSchema
 
 router = APIRouter()
 
+# Initialize CRUD
+eval_crud = CRUDBase(EvalDataset)
+task_crud = CRUDBase(Task)
+
 
 class GenerateEvalRequest(BaseModel):
     """Request for generating evaluation dataset"""
-    name: str
-    question_type: str = "mixed"
-    count: int = 50
+    name: str = Field(..., min_length=1, max_length=255)
+    question_type: str = Field("mixed", pattern="^(mixed|fact|reasoning|summary)$")
+    count: int = Field(50, ge=1, le=500)
 
 
 class RunEvalRequest(BaseModel):
@@ -26,18 +34,34 @@ class RunEvalRequest(BaseModel):
     model_config_id: Optional[UUID] = None
 
 
-@router.get("/", response_model=dict)
-async def list_eval_datasets(project_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.get("", response_model=ApiResponse)
+async def list_eval_datasets(
+    project_id: UUID,
+    page: int = Query(1, ge=1),
+    page_size: int = Query(20, ge=1, le=100),
+    db: AsyncSession = Depends(get_db)
+):
     """List evaluation datasets"""
-    result = await db.execute(
-        select(EvalDataset).where(EvalDataset.project_id == project_id).order_by(EvalDataset.created_at.desc())
+    skip = (page - 1) * page_size
+    datasets, total = await eval_crud.get_multi(
+        db,
+        skip=skip,
+        limit=page_size,
+        filters={"project_id": project_id},
+        order_by="created_at",
+        descending=True
     )
-    datasets = result.scalars().all()
 
-    return {"datasets": [EvalDatasetResponse.model_validate(d) for d in datasets]}
+    dataset_responses = [EvalDatasetResponse.model_validate(d) for d in datasets]
+    return PaginatedResponse.ok(
+        items=dataset_responses,
+        page=page,
+        page_size=page_size,
+        total=total
+    )
 
 
-@router.post("/", response_model=dict)
+@router.post("", response_model=ApiResponse)
 async def create_eval_dataset(
     project_id: UUID,
     request: GenerateEvalRequest,
@@ -53,10 +77,27 @@ async def create_eval_dataset(
     await db.commit()
     await db.refresh(db_dataset)
 
-    return {"id": str(db_dataset.id)}
+    return ApiResponse.ok(
+        data={"id": str(db_dataset.id)},
+        message="Evaluation dataset created successfully"
+    )
 
 
-@router.post("/{eval_id}/evaluate", response_model=dict)
+@router.get("/{eval_id}", response_model=ApiResponse)
+async def get_eval_dataset(
+    project_id: UUID,
+    eval_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
+    """Get evaluation dataset by ID"""
+    dataset = await eval_crud.get(db, eval_id)
+    if not dataset or dataset.project_id != project_id:
+        raise NotFoundException("Evaluation Dataset", eval_id)
+
+    return ApiResponse.ok(data=EvalDatasetResponse.model_validate(dataset))
+
+
+@router.post("/{eval_id}/evaluate", response_model=ApiResponse)
 async def run_evaluation(
     project_id: UUID,
     eval_id: UUID,
@@ -65,12 +106,9 @@ async def run_evaluation(
 ):
     """Run evaluation on dataset"""
     # Check dataset exists
-    result = await db.execute(
-        select(EvalDataset).where(EvalDataset.id == eval_id, EvalDataset.project_id == project_id)
-    )
-    dataset = result.scalar_one_or_none()
-    if not dataset:
-        raise HTTPException(status_code=404, detail="Evaluation dataset not found")
+    dataset = await eval_crud.get(db, eval_id)
+    if not dataset or dataset.project_id != project_id:
+        raise NotFoundException("Evaluation Dataset", eval_id)
 
     # Create evaluation task
     task = Task(
@@ -82,19 +120,21 @@ async def run_evaluation(
     await db.commit()
     await db.refresh(task)
 
-    # TODO: Start evaluation in background
-
-    return {"task_id": str(task.id), "message": "Evaluation task started"}
-
-
-@router.get("/results", response_model=dict)
-async def get_eval_results(project_id: UUID, task_id: UUID, db: AsyncSession = Depends(get_db)):
-    """Get evaluation results"""
-    result = await db.execute(
-        select(Task).where(Task.id == task_id, Task.project_id == project_id)
+    return ApiResponse.ok(
+        data={"task_id": str(task.id)},
+        message="Evaluation task started"
     )
-    task = result.scalar_one_or_none()
-    if not task:
-        raise HTTPException(status_code=404, detail="Task not found")
 
-    return TaskResponse.model_validate(task)
+
+@router.get("/results", response_model=ApiResponse)
+async def get_eval_results(
+    project_id: UUID,
+    task_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
+    """Get evaluation results"""
+    task = await task_crud.get(db, task_id)
+    if not task or task.project_id != project_id:
+        raise NotFoundException("Task", task_id)
+
+    return ApiResponse.ok(data=TaskResponse.model_validate(task))

backend/app/api/v1/files/__init__.py

@@ -2,24 +2,48 @@
 Files API Router
 """
 import os
-import aiofiles
+import asyncio
 from pathlib import Path
-from typing import List
-from uuid import UUID
-from fastapi import APIRouter, Depends, HTTPException, UploadFile, File, Form
+from typing import Optional
+from uuid import UUID, uuid4
+from fastapi import APIRouter, Depends, UploadFile, File, Query
+from fastapi.responses import FileResponse, PlainTextResponse
 from sqlalchemy.ext.asyncio import AsyncSession
 from sqlalchemy import select
-from app.core.database import get_db
+
+from app.api.response import ApiResponse, PaginatedResponse
 from app.core.config import get_settings
-from app.models.models import File
-from app.schemas.base import FileResponse
+from app.core.database import get_db
+from app.core.exceptions import ValidationException, NotFoundException
+from app.core.crud import CRUDBase
+from app.core.logging import log_success, log_failure
+from app.models.models import File as FileModel
+from app.models.models import Chunk, Question
+from app.schemas.file import FileResponse, FileCreateSchema
+from markitdown import MarkItDown
 
 settings = get_settings()
 router = APIRouter()
 
-# Ensure upload directory exists
-UPLOAD_DIR = Path(settings.UPLOAD_DIR)
-UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
+# Initialize CRUD
+file_crud = CRUDBase(FileModel)
+
+# Initialize markitdown
+markitdown = MarkItDown()
+
+
+def get_project_raw_dir(project_id: str) -> Path:
+    """获取项目的 raw 文件目录"""
+    base_dir = Path("/data/code/YG-Datasets/data") / project_id / "raw"
+    base_dir.mkdir(parents=True, exist_ok=True)
+    return base_dir
+
+
+def get_project_ready_dir(project_id: str) -> Path:
+    """获取项目的 ready 文件目录（处理后的文件）"""
+    base_dir = Path("/data/code/YG-Datasets/data") / project_id / "ready"
+    base_dir.mkdir(parents=True, exist_ok=True)
+    return base_dir
 
 
 def get_file_type(filename: str) -> str:
@@ -40,71 +64,342 @@ def get_file_type(filename: str) -> str:
     return type_map.get(ext, 'txt')
 
 
-@router.post("/upload", response_model=dict)
+# Allowed file extensions
+ALLOWED_EXTENSIONS = {'pdf', 'docx', 'doc', 'xlsx', 'xls', 'csv', 'epub', 'md', 'txt'}
+
+
+def validate_file(filename: str, file_size: int) -> None:
+    """Validate file extension and size"""
+    ext = filename.rsplit('.', 1)[-1].lower() if '.' in filename else ''
+
+    if ext not in ALLOWED_EXTENSIONS:
+        raise ValidationException(
+            f"File type '{ext}' not allowed",
+            field="file"
+        )
+
+    if file_size > settings.MAX_FILE_SIZE:
+        raise ValidationException(
+            f"File size exceeds maximum allowed size of {settings.MAX_FILE_SIZE // (1024*1024)}MB",
+            field="file"
+        )
+
+
+async def save_file_async(file: UploadFile, destination: Path) -> None:
+    """Save uploaded file asynchronously"""
+    content = await file.read()
+    loop = asyncio.get_event_loop()
+    await loop.run_in_executor(None, lambda: destination.write_bytes(content))
+
+
+@router.post("/upload", response_model=ApiResponse)
 async def upload_file(
     project_id: UUID,
     file: UploadFile = File(...),
     db: AsyncSession = Depends(get_db)
 ):
     """Upload a file"""
-    # Save file to disk
-    file_path = UPLOAD_DIR / f"{project_id}_{file.filename}"
-    async with aiofiles.open(file_path, 'wb') as f:
+    try:
+        # Read file content for validation
         content = await file.read()
-        await f.write(content)
+        file_size = len(content)
+
+        # Validate file
+        validate_file(file.filename, file_size)
+
+        # Save file to disk - 使用项目 raw 目录
+        safe_filename = f"{uuid4().hex[:8]}_{file.filename}"
+        project_dir = get_project_raw_dir(str(project_id))
+        file_path = project_dir / safe_filename
+
+        # Write file asynchronously
+        await asyncio.get_event_loop().run_in_executor(
+            None,
+            lambda: file_path.write_bytes(content)
+        )
 
         # Create file record
-    db_file = File(
+        db_file = FileModel(
             project_id=project_id,
             filename=file.filename,
             file_type=get_file_type(file.filename),
             file_path=str(file_path),
-        size=len(content),
-        status="pending"
+            size=file_size,
+            status="processing"
         )
         db.add(db_file)
         await db.commit()
         await db.refresh(db_file)
 
-    return {"id": str(db_file.id), "filename": db_file.filename, "status": db_file.status}
+        # 异步处理文件：立即返回，不等待处理完成
+        async def process_file_async(file_id: UUID, file_path_obj: Path, file_type: str, filename: str, project_id_val: UUID):
+            """后台异步处理文件"""
+            from sqlalchemy.ext.asyncio import AsyncSession
+            from app.core.database import AsyncSessionLocal
+
+            async with AsyncSessionLocal() as processing_db:
+                try:
+                    # 重新获取文件记录
+                    file_record = await file_crud.get(processing_db, file_id)
+                    if not file_record:
+                        return
+
+                    # 支持 markitdown 转换的文件类型
+                    markitdown_types = ["pdf", "docx", "doc", "pptx", "ppt", "xlsx", "xls", "htm", "html"]
+                    text_content = ""
+
+                    if file_type in markitdown_types:
+                        # 使用 markitdown 转换为 markdown
+                        loop = asyncio.get_event_loop()
+                        result = await loop.run_in_executor(
+                            None,
+                            lambda: markitdown.convert(str(file_path_obj))
+                        )
+                        text_content = result.text_content
+                    else:
+                        # txt, md 等直接读取
+                        text_content = file_path_obj.read_text(encoding='utf-8')
+
+                    # 保存到 ready 目录，使用 {uuid}.md 格式
+                    ready_dir = get_project_ready_dir(str(project_id_val))
+                    ready_filename = f"{file_id}.md"
+                    ready_path = ready_dir / ready_filename
+                    ready_path.write_text(text_content, encoding='utf-8')
+
+                    # 更新文件状态为处理完成
+                    file_record.status = "completed"
+                    await processing_db.commit()
+
+                    log_success(
+                        "文件处理完成",
+                        project_id=str(project_id_val),
+                        file_id=str(file_id),
+                        filename=filename,
+                        ready_path=str(ready_path)
+                    )
+
+                except Exception as e:
+                    # 更新文件状态为处理失败
+                    file_record = await file_crud.get(processing_db, file_id)
+                    if file_record:
+                        file_record.status = "failed"
+                        await processing_db.commit()
+
+                    log_failure(
+                        "文件处理失败",
+                        project_id=str(project_id_val),
+                        file_id=str(file_id),
+                        filename=filename,
+                        error=str(e)
+                    )
+
+        # 启动异步任务处理文件
+        asyncio.create_task(
+            process_file_async(
+                db_file.id,
+                file_path,
+                db_file.file_type,
+                file.filename,
+                project_id
+            )
+        )
+
+        return ApiResponse.ok(
+            data={"id": str(db_file.id), "filename": db_file.filename, "status": db_file.status},
+            message="File uploaded successfully, processing in background"
+        )
+    except Exception as e:
+        # 记录失败日志
+        log_failure(
+            "文件上传失败",
+            project_id=str(project_id),
+            filename=file.filename if 'file' in locals() else "unknown",
+            error=str(e)
+        )
+        raise
 
 
-@router.get("/", response_model=dict)
-async def list_files(project_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.get("", response_model=ApiResponse)
+async def list_files(
+    project_id: UUID,
+    page: int = Query(1, ge=1),
+    page_size: int = Query(20, ge=1, le=100),
+    db: AsyncSession = Depends(get_db)
+):
     """List files for a project"""
-    result = await db.execute(
-        select(File).where(File.project_id == project_id).order_by(File.created_at.desc())
+    skip = (page - 1) * page_size
+    files, total = await file_crud.get_multi(
+        db,
+        skip=skip,
+        limit=page_size,
+        filters={"project_id": project_id},
+        order_by="created_at",
+        descending=True
+    )
+
+    file_responses = [FileResponse.model_validate(f) for f in files]
+    return PaginatedResponse.ok(
+        items=file_responses,
+        page=page,
+        page_size=page_size,
+        total=total
     )
-    files = result.scalars().all()
-    return {"files": [FileResponse.model_validate(f) for f in files]}
 
 
-@router.get("/{file_id}", response_model=dict)
-async def get_file(project_id: UUID, file_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.get("/{file_id}", response_model=ApiResponse)
+async def get_file(
+    project_id: UUID,
+    file_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
     """Get file by ID"""
-    result = await db.execute(
-        select(File).where(File.id == file_id, File.project_id == project_id)
+    file = await file_crud.get(db, file_id)
+    if not file or file.project_id != project_id:
+        raise NotFoundException("File", file_id)
+
+    return ApiResponse.ok(data=FileResponse.model_validate(file))
+
+
+@router.get("/{file_id}/raw")
+async def get_file_raw(
+    project_id: UUID,
+    file_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
+    """Get raw file content for preview"""
+    file = await file_crud.get(db, file_id)
+    if not file or file.project_id != project_id:
+        raise NotFoundException("File", file_id)
+
+    # 读取 raw 目录中的原始文件
+    raw_path = Path(file.file_path)
+
+    if not raw_path.exists():
+        raise NotFoundException("File not found on disk", file_id)
+
+    # 根据文件类型返回不同的内容
+    if file.file_type in ['txt', 'md', 'markdown', 'csv']:
+        content = raw_path.read_text(encoding='utf-8')
+        return PlainTextResponse(content=content, media_type="text/plain; charset=utf-8")
+    elif file.file_type == 'pdf':
+        # 返回PDF文件，浏览器可以内嵌显示
+        import base64
+        content = raw_path.read_bytes()
+        b64 = base64.b64encode(content).decode('utf-8')
+        return PlainTextResponse(
+            content=f"data:application/pdf;base64,{b64}",
+            media_type="text/plain"
         )
-    file = result.scalar_one_or_none()
-    if not file:
-        raise HTTPException(status_code=404, detail="File not found")
-    return FileResponse.model_validate(file)
+    else:
+        # 其他二进制文件，返回文件信息
+        size_mb = file.size / (1024 * 1024)
+        content = f"""[二进制文件]
+
+文件名: {file.filename}
+文件类型: {file.file_type.upper()}
+文件大小: {size_mb:.2f} MB
+
+此文件为二进制格式，请下载后查看。
+"""
+        return PlainTextResponse(content=content, media_type="text/plain; charset=utf-8")
 
 
-@router.delete("/{file_id}", response_model=dict)
-async def delete_file(project_id: UUID, file_id: UUID, db: AsyncSession = Depends(get_db)):
-    """Delete file"""
-    result = await db.execute(
-        select(File).where(File.id == file_id, File.project_id == project_id)
+@router.get("/{file_id}/content")
+async def get_file_content(
+    project_id: UUID,
+    file_id: UUID,
+    db: AsyncSession = Depends(get_db)
+) -> PlainTextResponse:
+    """Get file content (markdown)"""
+    file = await file_crud.get(db, file_id)
+    if not file or file.project_id != project_id:
+        raise NotFoundException("File", file_id)
+
+    # 读取 ready 目录中的 markdown 文件
+    ready_path = Path("/data/code/YG-Datasets/data") / str(project_id) / "ready" / f"{file_id}.md"
+
+    if ready_path.exists():
+        content = ready_path.read_text(encoding='utf-8')
+        return PlainTextResponse(content=content, media_type="text/plain; charset=utf-8")
+    else:
+        raise NotFoundException("File content", file_id)
+
+
+@router.delete("/{file_id}", response_model=ApiResponse)
+async def delete_file(
+    project_id: UUID,
+    file_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
+    """Delete file and all related data (markdown, chunks, questions)"""
+    file = await file_crud.get(db, file_id)
+    if not file or file.project_id != project_id:
+        raise NotFoundException("File", file_id)
+
+    # Delete related chunks and their questions (explicit deletion for safety)
+    chunks_result = await db.execute(
+        select(Chunk).where(Chunk.file_id == file_id)
     )
-    file = result.scalar_one_or_none()
-    if not file:
-        raise HTTPException(status_code=404, detail="File not found")
+    chunks = chunks_result.scalars().all()
+    for chunk in chunks:
+        # Delete questions related to this chunk
+        questions_result = await db.execute(
+            select(Question).where(Question.chunk_id == chunk.id)
+        )
+        questions = questions_result.scalars().all()
+        for question in questions:
+            await db.delete(question)
+        # Delete chunk
+        await db.delete(chunk)
 
-    # Delete file from disk
+    # Delete file from raw directory
     if file.file_path and os.path.exists(file.file_path):
-        os.remove(file.file_path)
+        await asyncio.get_event_loop().run_in_executor(
+            None,
+            os.remove,
+            file.file_path
+        )
 
-    await db.delete(file)
+    # Delete file from ready directory (processed markdown) - try both naming conventions
+    ready_dir = Path("/data/code/YG-Datasets/data") / str(project_id) / "ready"
+    if ready_dir.exists():
+        # Try file_id.md (from upload process)
+        ready_path = ready_dir / f"{file_id}.md"
+        if ready_path.exists():
+            await asyncio.get_event_loop().run_in_executor(
+                None,
+                os.remove,
+                str(ready_path)
+            )
+        # Try file_id_filename.md (from split process)
+        for md_file in ready_dir.glob(f"{file_id}_*.md"):
+            await asyncio.get_event_loop().run_in_executor(
+                None,
+                os.remove,
+                str(md_file)
+            )
+
+    await file_crud.delete(db, file_id)
     await db.commit()
-    return {"message": "File deleted successfully"}
+    return ApiResponse.ok(message="File deleted successfully")
+
+
+@router.get("/{file_id}/download")
+async def download_file(
+    project_id: UUID,
+    file_id: UUID,
+    db: AsyncSession = Depends(get_db)
+) -> FileResponse:
+    """Download file"""
+    file = await file_crud.get(db, file_id)
+    if not file or file.project_id != project_id:
+        raise NotFoundException("File", file_id)
+
+    if not file.file_path or not os.path.exists(file.file_path):
+        raise ValidationException("File not found on disk", field="file")
+
+    return FileResponse(
+        path=file.file_path,
+        filename=file.filename,
+        media_type=f"application/{file.file_type}"
+    )

backend/app/api/v1/models/__init__.py

@@ -0,0 +1,305 @@
+"""
+Model API Router
+"""
+import uuid
+import httpx
+from fastapi import APIRouter, Depends, HTTPException
+from sqlalchemy.ext.asyncio import AsyncSession
+from sqlalchemy import select, update
+
+from app.core.database import get_db
+from app.api.response import ApiResponse
+from app.models.models import ModelConfig
+from app.schemas.model import ModelCreate, ModelUpdate, ModelResponse
+
+router = APIRouter()
+
+VALID_MODEL_TYPES = {"chat", "vlm", "embedding", "rerank"}
+
+
+def normalize_model_type(model_type: str | None, model_name: str | None) -> str:
+    """Normalize model type, with keyword fallback for legacy records."""
+    if model_type in VALID_MODEL_TYPES and model_type != "chat":
+        return model_type
+
+    normalized_name = (model_name or "").strip().lower()
+
+    rerank_keywords = ("rerank", "bce-reranker", "gte-rerank")
+    embedding_keywords = (
+        "embedding",
+        "embed",
+        "text-embedding",
+        "bge-",
+        "bge_m3",
+        "gte-",
+        "m3e",
+        "e5-",
+        "jina-embeddings",
+    )
+    vlm_keywords = ("vl", "vision", "visual", "multimodal", "qwen-vl", "gpt-4o")
+
+    if any(keyword in normalized_name for keyword in rerank_keywords):
+        return "rerank"
+    if any(keyword in normalized_name for keyword in embedding_keywords):
+        return "embedding"
+    if any(keyword in normalized_name for keyword in vlm_keywords):
+        return "vlm"
+    return model_type if model_type in VALID_MODEL_TYPES else "chat"
+
+
+async def test_model_connection(model: ModelConfig) -> dict:
+    """Test model connection by calling the API"""
+    if not model.api_key:
+        return {"success": False, "message": "API Key is missing"}
+
+    api_base = model.api_base or ""
+    provider = model.provider
+    model_name = model.model_name
+    model_type = normalize_model_type(model.model_type, model_name)
+    api_key = model.api_key
+
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json"
+    }
+
+    try:
+        async with httpx.AsyncClient(timeout=10.0) as client:
+            if model_type in {"chat", "vlm"} and provider in {"openai", "ali"}:
+                # OpenAI compatible API test
+                response = await client.post(
+                    f"{api_base.rstrip('/')}/chat/completions",
+                    headers=headers,
+                    json={
+                        "model": model_name,
+                        "messages": [{"role": "user", "content": "Hi"}],
+                        "max_tokens": 5
+                    }
+                )
+            elif model_type in {"chat", "vlm"} and provider == "minimax":
+                # MiniMax API test
+                response = await client.post(
+                    f"{api_base.rstrip('/')}/chat/completions_v2",
+                    headers={
+                        **headers,
+                        "Authorization": f"Bearer {api_key}"
+                    },
+                    json={
+                        "model": model_name,
+                        "messages": [{"role": "user", "content": "Hi"}]
+                    }
+                )
+            elif model_type in {"chat", "vlm"} and provider == "glm":
+                # GLM API test
+                response = await client.post(
+                    f"{api_base.rstrip('/')}/chat/completions",
+                    headers=headers,
+                    json={
+                        "model": model_name,
+                        "messages": [{"role": "user", "content": "Hi"}]
+                    }
+                )
+            elif model_type == "embedding" and provider in {"openai", "ali", "glm"}:
+                response = await client.post(
+                    f"{api_base.rstrip('/')}/embeddings",
+                    headers=headers,
+                    json={
+                        "model": model_name,
+                        "input": "test"
+                    }
+                )
+            elif model_type == "embedding" and provider == "minimax":
+                return {"success": False, "message": "MiniMax embedding 自动测试暂未接入，请手动确认端点与模型"}
+            elif model_type == "rerank":
+                return {"success": False, "message": "Rerank 自动测试暂未接入，请先保存配置并在实际流程中验证"}
+            else:
+                return {"success": False, "message": f"Unsupported provider/type: {provider}/{model_type}"}
+
+            if response.status_code == 200:
+                return {"success": True, "message": "Connection successful"}
+            else:
+                return {"success": False, "message": f"API error: {response.status_code} - {response.text[:100]}"}
+
+    except httpx.TimeoutException:
+        return {"success": False, "message": "Connection timeout"}
+    except Exception as e:
+        return {"success": False, "message": f"Connection failed: {str(e)}"}
+
+
+# Helper to convert string to UUID
+def parse_uuid(id_str: str) -> uuid.UUID:
+    """Parse string to UUID"""
+    try:
+        return uuid.UUID(id_str)
+    except ValueError:
+        raise HTTPException(status_code=400, detail="Invalid UUID format")
+
+
+@router.get("", response_model=ApiResponse)
+async def list_models(db: AsyncSession = Depends(get_db)):
+    """Get all models"""
+    result = await db.execute(
+        select(ModelConfig).where(ModelConfig.project_id == None)  # noqa: E711
+    )
+    models = result.scalars().all()
+    # Convert to Pydantic schema
+    model_responses = [ModelResponse.model_validate(m) for m in models]
+    return ApiResponse(data=model_responses)
+
+
+@router.post("", response_model=ApiResponse)
+async def create_model(model: ModelCreate, db: AsyncSession = Depends(get_db)):
+    """Create a new model"""
+    # If setting as default, unset other defaults first
+    if model.is_default == "true":
+        await db.execute(
+            update(ModelConfig)
+            .where(ModelConfig.project_id == None)  # noqa: E711
+            .values(is_default="false")
+        )
+
+    db_model = ModelConfig(
+        provider=model.provider,
+        model_type=model.model_type,
+        model_name=model.model_name,
+        api_key=model.api_key,
+        api_base=model.api_base,
+        is_default=model.is_default,
+        project_id=None  # Global model config
+    )
+    db.add(db_model)
+    await db.commit()
+    await db.refresh(db_model)
+    # Convert to Pydantic schema
+    response = ModelResponse.model_validate(db_model)
+    return ApiResponse(data=response)
+
+
+@router.get("/{model_id}", response_model=ApiResponse)
+async def get_model(model_id: str, db: AsyncSession = Depends(get_db)):
+    """Get a model by ID"""
+    model_uuid = parse_uuid(model_id)
+    result = await db.execute(
+        select(ModelConfig).where(
+            ModelConfig.id == model_uuid,
+            ModelConfig.project_id == None  # noqa: E711
+        )
+    )
+    model = result.scalar_one_or_none()
+    if not model:
+        raise HTTPException(status_code=404, detail="Model not found")
+    response = ModelResponse.model_validate(model)
+    return ApiResponse(data=response)
+
+
+@router.put("/{model_id}", response_model=ApiResponse)
+async def update_model(model_id: str, model_update: ModelUpdate, db: AsyncSession = Depends(get_db)):
+    """Update a model"""
+    model_uuid = parse_uuid(model_id)
+    result = await db.execute(
+        select(ModelConfig).where(
+            ModelConfig.id == model_uuid,
+            ModelConfig.project_id == None  # noqa: E711
+        )
+    )
+    model = result.scalar_one_or_none()
+    if not model:
+        raise HTTPException(status_code=404, detail="Model not found")
+
+    # If setting as default, unset other defaults first
+    if model_update.is_default == "true":
+        await db.execute(
+            update(ModelConfig)
+            .where(
+                ModelConfig.project_id == None,  # noqa: E711
+                ModelConfig.id != model_uuid
+            )
+            .values(is_default="false")
+        )
+
+    update_data = model_update.model_dump(exclude_unset=True)
+    for key, value in update_data.items():
+        setattr(model, key, value)
+
+    await db.commit()
+    await db.refresh(model)
+    response = ModelResponse.model_validate(model)
+    return ApiResponse(data=response)
+
+
+@router.delete("/{model_id}", response_model=ApiResponse)
+async def delete_model(model_id: str, db: AsyncSession = Depends(get_db)):
+    """Delete a model"""
+    model_uuid = parse_uuid(model_id)
+    result = await db.execute(
+        select(ModelConfig).where(
+            ModelConfig.id == model_uuid,
+            ModelConfig.project_id == None  # noqa: E711
+        )
+    )
+    model = result.scalar_one_or_none()
+    if not model:
+        raise HTTPException(status_code=404, detail="Model not found")
+
+    await db.delete(model)
+    await db.commit()
+    return ApiResponse(message="Model deleted successfully")
+
+
+@router.post("/{model_id}/set-default", response_model=ApiResponse)
+async def set_default_model(model_id: str, db: AsyncSession = Depends(get_db)):
+    """Set a model as default"""
+    model_uuid = parse_uuid(model_id)
+    result = await db.execute(
+        select(ModelConfig).where(
+            ModelConfig.id == model_uuid,
+            ModelConfig.project_id == None  # noqa: E711
+        )
+    )
+    model = result.scalar_one_or_none()
+    if not model:
+        raise HTTPException(status_code=404, detail="Model not found")
+
+    # Unset all other defaults
+    await db.execute(
+        update(ModelConfig)
+        .where(
+            ModelConfig.project_id == None,  # noqa: E711
+            ModelConfig.id != model_uuid
+        )
+        .values(is_default="false")
+    )
+
+    model.is_default = "true"
+    await db.commit()
+    await db.refresh(model)
+    response = ModelResponse.model_validate(model)
+    return ApiResponse(data=response)
+
+
+@router.post("/{model_id}/test", response_model=ApiResponse)
+async def test_model(model_id: str, db: AsyncSession = Depends(get_db)):
+    """Test model connection"""
+    model_uuid = parse_uuid(model_id)
+    result = await db.execute(
+        select(ModelConfig).where(
+            ModelConfig.id == model_uuid,
+            ModelConfig.project_id == None  # noqa: E711
+        )
+    )
+    model = result.scalar_one_or_none()
+    if not model:
+        raise HTTPException(status_code=404, detail="Model not found")
+
+    # Test the connection
+    test_result = await test_model_connection(model)
+
+    # Save connection status to database
+    model.model_type = normalize_model_type(model.model_type, model.model_name)
+    model.connection_status = "connected" if test_result["success"] else "disconnected"
+    await db.commit()
+    await db.refresh(model)
+
+    # Return updated model
+    response = ModelResponse.model_validate(model)
+    return ApiResponse(data={"test_result": test_result, "model": response})

backend/app/api/v1/projects/__init__.py

@@ -1,74 +1,120 @@
 """
 Projects API Router
 """
-from typing import List
+import logging
+import shutil
+from pathlib import Path
+from typing import List, Optional
 from uuid import UUID
-from fastapi import APIRouter, Depends, HTTPException
+from fastapi import APIRouter, Depends, Query
 from sqlalchemy.ext.asyncio import AsyncSession
-from sqlalchemy import select
+
+from app.api.response import ApiResponse, PaginatedResponse
 from app.core.database import get_db
+from app.core.exceptions import NotFoundException
+from app.core.crud import CRUDBase
 from app.models.models import Project
-from app.schemas.base import (
+from app.schemas.project import (
     ProjectCreate,
     ProjectUpdate,
-    ProjectResponse
+    ProjectResponse,
+    ProjectCreateSchema,
+    ProjectUpdateSchema
 )
 
 router = APIRouter()
+logger = logging.getLogger("yg_dataset.projects")
+
+# Initialize CRUD
+project_crud = CRUDBase(Project)
 
 
-@router.get("/", response_model=dict)
-async def list_projects(db: AsyncSession = Depends(get_db)):
-    """List all projects"""
-    result = await db.execute(select(Project).order_by(Project.created_at.desc()))
-    projects = result.scalars().all()
-    return {"projects": [ProjectResponse.model_validate(p) for p in projects]}
+@router.get("", response_model=PaginatedResponse)
+async def list_projects(
+    page: int = Query(1, ge=1, description="Page number"),
+    page_size: int = Query(20, ge=1, le=100, description="Page size"),
+    db: AsyncSession = Depends(get_db)
+):
+    """List all projects with pagination"""
+    logger.info(f"Listing projects - page: {page}, page_size: {page_size}")
+    skip = (page - 1) * page_size
+    projects, total = await project_crud.get_multi(
+        db,
+        skip=skip,
+        limit=page_size,
+        order_by="created_at",
+        descending=True
+    )
+
+    logger.info(f"Found {total} projects, returning {len(projects)} items")
+    project_responses = [ProjectResponse.model_validate(p) for p in projects]
+    return PaginatedResponse.ok(
+        items=project_responses,
+        page=page,
+        page_size=page_size,
+        total=total
+    )
 
 
-@router.post("/", response_model=dict)
-async def create_project(project: ProjectCreate, db: AsyncSession = Depends(get_db)):
+@router.post("", response_model=ApiResponse)
+async def create_project(
+    project: ProjectCreateSchema,
+    db: AsyncSession = Depends(get_db)
+):
     """Create a new project"""
-    db_project = Project(**project.model_dump())
-    db.add(db_project)
-    await db.commit()
-    await db.refresh(db_project)
-    return {"id": str(db_project.id)}
+    logger.info(f"Creating project: name={project.name}, description={project.description}")
+    db_project = await project_crud.create(db, project)
+    logger.info(f"Project created successfully: id={db_project.id}")
+    return ApiResponse.ok(
+        data={"id": str(db_project.id)},
+        message="Project created successfully"
+    )
 
 
-@router.get("/{project_id}", response_model=dict)
-async def get_project(project_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.get("/{project_id}", response_model=ApiResponse)
+async def get_project(
+    project_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
     """Get project by ID"""
-    result = await db.execute(select(Project).where(Project.id == project_id))
-    project = result.scalar_one_or_none()
-    if not project:
-        raise HTTPException(status_code=404, detail="Project not found")
-    return ProjectResponse.model_validate(project)
+    logger.info(f"Getting project: id={project_id}")
+    project = await project_crud.get_or_raise(db, project_id, "Project")
+    logger.info(f"Found project: name={project.name}")
+    return ApiResponse.ok(data=ProjectResponse.model_validate(project))
 
 
-@router.put("/{project_id}", response_model=dict)
-async def update_project(project_id: UUID, project: ProjectUpdate, db: AsyncSession = Depends(get_db)):
+@router.put("/{project_id}", response_model=ApiResponse)
+async def update_project(
+    project_id: UUID,
+    project: ProjectUpdateSchema,
+    db: AsyncSession = Depends(get_db)
+):
     """Update project"""
-    result = await db.execute(select(Project).where(Project.id == project_id))
-    db_project = result.scalar_one_or_none()
-    if not db_project:
-        raise HTTPException(status_code=404, detail="Project not found")
-
-    for key, value in project.model_dump(exclude_unset=True).items():
-        setattr(db_project, key, value)
-
-    await db.commit()
-    await db.refresh(db_project)
-    return ProjectResponse.model_validate(db_project)
+    logger.info(f"Updating project: id={project_id}")
+    db_project = await project_crud.get_or_raise(db, project_id, "Project")
+    updated_project = await project_crud.update(db, db_project, project)
+    logger.info(f"Project updated: name={updated_project.name}")
+    return ApiResponse.ok(
+        data=ProjectResponse.model_validate(updated_project),
+        message="Project updated successfully"
+    )
 
 
-@router.delete("/{project_id}", response_model=dict)
-async def delete_project(project_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.delete("/{project_id}", response_model=ApiResponse)
+async def delete_project(
+    project_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
     """Delete project"""
-    result = await db.execute(select(Project).where(Project.id == project_id))
-    project = result.scalar_one_or_none()
-    if not project:
-        raise HTTPException(status_code=404, detail="Project not found")
+    logger.info(f"Deleting project: id={project_id}")
+    await project_crud.get_or_raise(db, project_id, "Project")
+    await project_crud.delete(db, project_id)
 
-    await db.delete(project)
-    await db.commit()
-    return {"message": "Project deleted successfully"}
+    # 删除项目对应的本地数据目录
+    project_data_dir = Path("/data/code/YG-Datasets/data") / str(project_id)
+    if project_data_dir.exists():
+        shutil.rmtree(project_data_dir)
+        logger.info(f"Project data directory deleted: {project_data_dir}")
+
+    logger.info(f"Project deleted: id={project_id}")
+    return ApiResponse.ok(message="Project deleted successfully")

backend/app/api/v1/questions/__init__.py

@@ -1,122 +1,402 @@
 """
 Questions API Router
 """
+import asyncio
+import json
+import re
 from typing import List, Optional
 from uuid import UUID
-from pydantic import BaseModel
-from fastapi import APIRouter, Depends, HTTPException, Query
-from sqlalchemy.ext.asyncio import AsyncSession
+
+import httpx
+from fastapi import APIRouter, Depends, Query
+from pydantic import BaseModel, Field
 from sqlalchemy import select
-from app.core.database import get_db
-from app.models.models import Question, Chunk
-from app.schemas.base import QuestionCreate, QuestionResponse
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.api.response import ApiResponse, PaginatedResponse
+from app.core.crud import CRUDBase
+from app.core.database import AsyncSessionLocal, get_db
+from app.core.exceptions import NotFoundException, ValidationException
+from app.core.logging import log_failure, log_success
+from app.models.models import Chunk, ModelConfig, Question
+from app.schemas.question import QuestionCreateSchema, QuestionResponse
 
 router = APIRouter()
 
+# Initialize CRUD
+question_crud = CRUDBase(Question)
+
+VALID_MODEL_TYPES = {"chat", "vlm", "embedding", "rerank"}
+
+DEFAULT_PRESET_PROMPT = (
+    "你是一名高质量中文问答数据构建助手。"
+    "请基于给定 chunk 内容生成准确、自然、可用于训练的数据集问答对。"
+    "问题必须清晰具体，答案必须直接来自内容或基于内容做合理概括，"
+    "不要编造原文没有的信息，不要输出与目录、导航、页眉页脚、噪声文字相关的问题。"
+)
+
 
 class GenerateRequest(BaseModel):
     """Request model for generating questions"""
-    chunk_ids: List[UUID] = []
-    count: int = 5
-    question_types: List[str] = ["fact", "summary"]
+    chunk_ids: List[UUID] = Field(..., min_length=1)
+    model_id: UUID
+    count: int = Field(3, ge=1, le=10)
+    dirty_data_filter: bool = True
+    thinking_mode: bool = True
+    preset_prompt: str = Field(default=DEFAULT_PRESET_PROMPT, min_length=1, max_length=4000)
 
 
-@router.post("/generate", response_model=dict)
+def normalize_model_type(model_type: str | None, model_name: str | None) -> str:
+    """Normalize model type, with keyword fallback for legacy records."""
+    if model_type in VALID_MODEL_TYPES and model_type != "chat":
+        return model_type
+
+    normalized_name = (model_name or "").strip().lower()
+    rerank_keywords = ("rerank", "bce-reranker", "gte-rerank")
+    embedding_keywords = (
+        "embedding",
+        "embed",
+        "text-embedding",
+        "bge-",
+        "bge_m3",
+        "gte-",
+        "m3e",
+        "e5-",
+        "jina-embeddings",
+    )
+    vlm_keywords = ("vl", "vision", "visual", "multimodal", "qwen-vl", "gpt-4o")
+
+    if any(keyword in normalized_name for keyword in rerank_keywords):
+        return "rerank"
+    if any(keyword in normalized_name for keyword in embedding_keywords):
+        return "embedding"
+    if any(keyword in normalized_name for keyword in vlm_keywords):
+        return "vlm"
+    return model_type if model_type in VALID_MODEL_TYPES else "chat"
+
+
+def is_dirty_chunk(content: str) -> bool:
+    """Heuristic dirty-data filter for low-value chunks."""
+    normalized = re.sub(r"\s+", " ", (content or "")).strip()
+    if len(normalized) < 40:
+        return True
+
+    if len(re.sub(r"[^\u4e00-\u9fffA-Za-z0-9]", "", normalized)) < 24:
+        return True
+
+    lowered = normalized.lower()
+    if lowered in {"目录", "contents", "table of contents"}:
+        return True
+
+    lines = [line.strip() for line in (content or "").splitlines() if line.strip()]
+    if lines:
+        short_lines = sum(1 for line in lines if len(line) <= 18)
+        dotted_lines = sum(1 for line in lines if re.search(r"[·•…\.]{3,}|\s\d+$", line))
+        if short_lines / len(lines) > 0.7 and len(lines) >= 3:
+            return True
+        if dotted_lines / len(lines) > 0.4:
+            return True
+
+    punctuation_ratio = sum(1 for ch in normalized if not ch.isalnum() and not ("\u4e00" <= ch <= "\u9fff")) / max(len(normalized), 1)
+    if punctuation_ratio > 0.45:
+        return True
+
+    return False
+
+
+def build_generation_prompt(chunk: Chunk, request: GenerateRequest) -> str:
+    """Build user prompt for QA generation."""
+    thinking_instruction = (
+        "请先对内容做简短分析，识别核心事实、概念、关系与潜在考点，然后再生成问答。"
+        "分析过程只用于提高质量，不要在最终输出中暴露你的思维链。"
+        if request.thinking_mode
+        else "直接基于内容生成高质量问答。"
+    )
+
+    return (
+        f"{request.preset_prompt}\n\n"
+        "输出要求：\n"
+        f"1. 生成 {request.count} 组问答。\n"
+        "2. 只输出 JSON 数组，不要输出解释、标题、Markdown。\n"
+        '3. 每个对象结构为 {"question":"...","answer":"...","question_type":"fact|summary|reasoning"}。\n'
+        "4. 问题避免重复，答案避免空泛。\n"
+        "5. 如果内容不足以生成高质量问答，请返回空数组 []。\n"
+        f"6. {thinking_instruction}\n\n"
+        f"Chunk 名称：{chunk.name or '未命名分片'}\n"
+        f"Chunk 内容：\n{chunk.content}"
+    )
+
+
+def extract_text_from_response(data: dict) -> str:
+    """Extract response text from provider response."""
+    choices = data.get("choices") or []
+    if choices:
+        message = choices[0].get("message") or {}
+        content = message.get("content")
+        if isinstance(content, str):
+            return content
+        if isinstance(content, list):
+            parts = [item.get("text", "") for item in content if isinstance(item, dict)]
+            return "\n".join(part for part in parts if part)
+    return ""
+
+
+def parse_generated_questions(raw_text: str) -> List[dict]:
+    """Parse JSON array from model output."""
+    text = (raw_text or "").strip()
+    if not text:
+        return []
+
+    fenced_match = re.search(r"```json\s*(.*?)\s*```", text, flags=re.S)
+    if fenced_match:
+        text = fenced_match.group(1).strip()
+
+    if not text.startswith("["):
+        array_match = re.search(r"(\[\s*\{.*\}\s*\])", text, flags=re.S)
+        if array_match:
+            text = array_match.group(1)
+
+    try:
+        parsed = json.loads(text)
+    except json.JSONDecodeError:
+        return []
+
+    if not isinstance(parsed, list):
+        return []
+
+    normalized = []
+    for item in parsed:
+        if not isinstance(item, dict):
+            continue
+        question = str(item.get("question", "")).strip()
+        answer = str(item.get("answer", "")).strip()
+        question_type = str(item.get("question_type", "fact")).strip() or "fact"
+        if not question or not answer:
+            continue
+        normalized.append({
+            "question": question,
+            "answer": answer,
+            "question_type": question_type
+        })
+    return normalized
+
+
+async def call_generation_model(model: ModelConfig, prompt: str) -> str:
+    """Call configured chat model for question generation."""
+    provider = model.provider
+    api_base = (model.api_base or "").rstrip("/")
+    api_key = model.api_key or ""
+    model_name = model.model_name
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json"
+    }
+    payload = {
+        "model": model_name,
+        "messages": [
+            {
+                "role": "system",
+                "content": "你是问答数据构建助手。严格按 JSON 输出，不要输出额外说明。"
+            },
+            {
+                "role": "user",
+                "content": prompt
+            }
+        ],
+        "temperature": 0.4,
+        "response_format": {"type": "json_object"}
+    }
+
+    async with httpx.AsyncClient(timeout=120.0) as client:
+        if provider == "minimax":
+            response = await client.post(
+                f"{api_base}/chat/completions_v2",
+                headers=headers,
+                json={k: v for k, v in payload.items() if k != "response_format"}
+            )
+        else:
+            response = await client.post(
+                f"{api_base}/chat/completions",
+                headers=headers,
+                json=payload
+            )
+
+        response.raise_for_status()
+        data = response.json()
+        content = extract_text_from_response(data)
+        if not content:
+            raise ValueError("Model returned empty content")
+
+        if content.lstrip().startswith("{"):
+            obj = json.loads(content)
+            if isinstance(obj, dict) and isinstance(obj.get("questions"), list):
+                return json.dumps(obj["questions"], ensure_ascii=False)
+        return content
+
+
+async def process_generate_async(project_id: UUID, request: GenerateRequest):
+    """Generate QA pairs in background."""
+    async with AsyncSessionLocal() as db:
+        try:
+            model_result = await db.execute(
+                select(ModelConfig).where(ModelConfig.id == request.model_id, ModelConfig.project_id == None)  # noqa: E711
+            )
+            model = model_result.scalar_one_or_none()
+            if not model:
+                return
+
+            model_type = normalize_model_type(model.model_type, model.model_name)
+            if model_type not in {"chat", "vlm"}:
+                raise ValidationException("Selected model must be chat/vlm type", field="model_id")
+
+            chunk_result = await db.execute(
+                select(Chunk).where(Chunk.id.in_(request.chunk_ids), Chunk.project_id == project_id)
+            )
+            chunks = chunk_result.scalars().all()
+            if not chunks:
+                return
+
+            created_count = 0
+            skipped_count = 0
+            for chunk in chunks:
+                if request.dirty_data_filter and is_dirty_chunk(chunk.content):
+                    skipped_count += 1
+                    continue
+
+                prompt = build_generation_prompt(chunk, request)
+                raw_text = await call_generation_model(model, prompt)
+                qa_pairs = parse_generated_questions(raw_text)[:request.count]
+                if not qa_pairs:
+                    skipped_count += 1
+                    continue
+
+                for item in qa_pairs:
+                    db.add(Question(
+                        project_id=project_id,
+                        chunk_id=chunk.id,
+                        content=item["question"],
+                        answer=item["answer"],
+                        question_type=item["question_type"],
+                        source="generated"
+                    ))
+                    created_count += 1
+
+                await db.commit()
+
+            log_success(
+                "问答批量生成完成",
+                project_id=str(project_id),
+                model_id=str(model.id),
+                chunk_count=len(chunks),
+                created_questions=created_count,
+                skipped_chunks=skipped_count
+            )
+        except Exception as e:
+            log_failure(
+                "问答批量生成失败",
+                project_id=str(project_id),
+                model_id=str(request.model_id),
+                error=str(e)
+            )
+
+
+@router.post("/generate", response_model=ApiResponse)
 async def generate_questions(
     project_id: UUID,
     request: GenerateRequest,
     db: AsyncSession = Depends(get_db)
 ):
-    """Generate questions from chunks using LLM"""
-    # TODO: Implement LLM-based question generation
-    # This is a placeholder that creates sample questions
-
-    if not request.chunk_ids:
-        raise HTTPException(status_code=400, detail="chunk_ids is required")
-
-    # Get chunks
-    result = await db.execute(
-        select(Chunk).where(Chunk.id.in_(request.chunk_ids), Chunk.project_id == project_id)
+    """Generate questions from chunks using LLM in background."""
+    model_result = await db.execute(
+        select(ModelConfig).where(ModelConfig.id == request.model_id, ModelConfig.project_id == None)  # noqa: E711
     )
-    chunks = result.scalars().all()
+    model = model_result.scalar_one_or_none()
+    if not model:
+        raise ValidationException("Selected model not found", field="model_id")
 
-    if not chunks:
-        raise HTTPException(status_code=404, detail="No chunks found")
+    model_type = normalize_model_type(model.model_type, model.model_name)
+    if model_type not in {"chat", "vlm"}:
+        raise ValidationException("Selected model must be chat/vlm type", field="model_id")
+    if not model.api_key:
+        raise ValidationException("Selected model is missing API Key", field="model_id")
 
-    # Create sample questions (placeholder)
-    created_questions = []
-    for chunk in chunks:
-        for i in range(request.count):
-            question = Question(
-                project_id=project_id,
-                chunk_id=chunk.id,
-                content=f"这是关于「{chunk.name}」的问题 {i+1}？",
-                answer=f"这是问题 {i+1} 的答案。",
-                question_type=request.question_types[0] if request.question_types else "fact",
-                source="generated"
+    chunk_result = await db.execute(
+        select(Chunk.id).where(Chunk.id.in_(request.chunk_ids), Chunk.project_id == project_id)
     )
-            db.add(question)
-            created_questions.append(question)
+    valid_chunk_ids = [row[0] for row in chunk_result.all()]
+    if not valid_chunk_ids:
+        raise ValidationException("No valid chunks found", field="chunk_ids")
 
-    await db.commit()
+    request_payload = request.model_copy(update={"chunk_ids": valid_chunk_ids})
+    asyncio.create_task(process_generate_async(project_id, request_payload))
 
-    return {
-        "questions": len(created_questions),
-        "message": f"Successfully generated {len(created_questions)} questions"
-    }
+    return ApiResponse.ok(
+        data={"chunk_count": len(valid_chunk_ids), "status": "processing"},
+        message="Question generation started in background"
+    )
 
 
-@router.get("/", response_model=dict)
+@router.get("", response_model=ApiResponse)
 async def list_questions(
     project_id: UUID,
     chunk_id: Optional[UUID] = Query(None),
+    page: int = Query(1, ge=1),
+    page_size: int = Query(20, ge=1, le=100),
     db: AsyncSession = Depends(get_db)
 ):
     """List questions for a project"""
-    query = select(Question).where(Question.project_id == project_id)
-
+    filters = {"project_id": project_id}
     if chunk_id:
-        query = query.where(Question.chunk_id == chunk_id)
+        filters["chunk_id"] = chunk_id
 
-    result = await db.execute(query)
-    questions = result.scalars().all()
+    skip = (page - 1) * page_size
+    questions, total = await question_crud.get_multi(
+        db,
+        skip=skip,
+        limit=page_size,
+        filters=filters,
+        order_by="created_at",
+        descending=True
+    )
 
-    return {"questions": [QuestionResponse.model_validate(q) for q in questions]}
+    question_responses = [QuestionResponse.model_validate(q) for q in questions]
+    return PaginatedResponse.ok(
+        items=question_responses,
+        page=page,
+        page_size=page_size,
+        total=total
+    )
 
 
-@router.put("/{question_id}", response_model=dict)
+@router.put("/{question_id}", response_model=ApiResponse)
 async def update_question(
     project_id: UUID,
     question_id: UUID,
-    question: QuestionCreate,
+    question: QuestionCreateSchema,
     db: AsyncSession = Depends(get_db)
 ):
     """Update question"""
-    result = await db.execute(
-        select(Question).where(Question.id == question_id, Question.project_id == project_id)
+    db_question = await question_crud.get(db, question_id)
+    if not db_question or db_question.project_id != project_id:
+        raise NotFoundException("Question", question_id)
+
+    updated_question = await question_crud.update(db, db_question, question)
+    return ApiResponse.ok(
+        data=QuestionResponse.model_validate(updated_question),
+        message="Question updated successfully"
     )
-    db_question = result.scalar_one_or_none()
-    if not db_question:
-        raise HTTPException(status_code=404, detail="Question not found")
-
-    for key, value in question.model_dump(exclude_unset=True).items():
-        setattr(db_question, key, value)
-
-    await db.commit()
-    await db.refresh(db_question)
-    return QuestionResponse.model_validate(db_question)
 
 
-@router.delete("/{question_id}", response_model=dict)
-async def delete_question(project_id: UUID, question_id: UUID, db: AsyncSession = Depends(get_db)):
+@router.delete("/{question_id}", response_model=ApiResponse)
+async def delete_question(
+    project_id: UUID,
+    question_id: UUID,
+    db: AsyncSession = Depends(get_db)
+):
     """Delete question"""
-    result = await db.execute(
-        select(Question).where(Question.id == question_id, Question.project_id == project_id)
-    )
-    question = result.scalar_one_or_none()
-    if not question:
-        raise HTTPException(status_code=404, detail="Question not found")
+    question = await question_crud.get(db, question_id)
+    if not question or question.project_id != project_id:
+        raise NotFoundException("Question", question_id)
 
-    await db.delete(question)
-    await db.commit()
-    return {"message": "Question deleted successfully"}
+    await question_crud.delete(db, question_id)
+    return ApiResponse.ok(message="Question deleted successfully")

backend/app/core/auth.py

@@ -0,0 +1,38 @@
+"""
+API Key Authentication
+API Key 认证中间件
+"""
+from typing import Optional
+from fastapi import Header, HTTPException, Request
+from fastapi.security import APIKeyHeader
+
+from app.core.config import get_settings
+
+settings = get_settings()
+
+# API Key header
+API_KEY_HEADER = APIKeyHeader(name="X-API-Key", auto_error=False)
+
+
+async def verify_api_key(api_key: Optional[str] = Header(None)) -> str:
+    """Verify API key from header"""
+    if not api_key:
+        raise HTTPException(status_code=401, detail="API key is required")
+
+    # In production, you would validate against a database or cache
+    # For development, we can use a simple validation
+    if settings.DEBUG and api_key == "dev-api-key":
+        return api_key
+
+    # TODO: Implement proper API key validation
+    # This is a placeholder - in production, validate against stored keys
+    if len(api_key) < 32:
+        raise HTTPException(status_code=401, detail="Invalid API key")
+
+    return api_key
+
+
+def create_api_key() -> str:
+    """Generate a new API key"""
+    import secrets
+    return secrets.token_hex(32)

backend/app/core/config.py

@@ -4,7 +4,7 @@ Application Configuration
 
 from functools import lru_cache
 from pydantic_settings import BaseSettings
-from pydantic import Field
+from pydantic import Field, field_validator
 
 
 class Settings(BaseSettings):
@@ -15,12 +15,16 @@ class Settings(BaseSettings):
     DEBUG: bool = True
     HOST: str = "0.0.0.0"
     PORT: int = 8000
+    ALLOWED_ORIGINS: str = Field(
+        default="*",
+        description="Comma-separated list of allowed CORS origins"
+    )
 
     # Database - 使用 SQLite 进行开发/测试
     # 生产环境可切换为 PostgreSQL
     DATABASE_URL: str = Field(
-        default="sqlite:///./ygdataset.db",
-        description="Database connection URL (sqlite:// or postgresql+asyncpg://)"
+        default="sqlite+aiosqlite:///./ygdataset.db",
+        description="Database connection URL (sqlite+aiosqlite:// or postgresql+asyncpg://)"
     )
     DATABASE_URL_SYNC: str = Field(
         default="sqlite:///./ygdataset.db",
@@ -38,8 +42,31 @@ class Settings(BaseSettings):
     DEFAULT_MODEL_PROVIDER: str = "openai"
     DEFAULT_MODEL_NAME: str = "gpt-4o-mini"
 
+    # Security
+    SECRET_KEY: str = Field(
+        default="your-secret-key-change-in-production",
+        description="Secret key for JWT and other security operations"
+    )
+    API_KEY_HEADER: str = "X-API-Key"
+
+    # Pagination
+    DEFAULT_PAGE_SIZE: int = 20
+    MAX_PAGE_SIZE: int = 100
+
+    # Logging
+    LOG_LEVEL: str = "INFO"
+
+    @field_validator("MAX_FILE_SIZE")
+    @classmethod
+    def validate_max_file_size(cls, v: int) -> int:
+        """Validate max file size (max 500MB)"""
+        if v > 500 * 1024 * 1024:
+            return 500 * 1024 * 1024
+        return v
+
     class Config:
         env_file = ".env"
+        env_file_encoding = "utf-8"
         extra = "allow"
 
 
@@ -47,3 +74,7 @@ class Settings(BaseSettings):
 def get_settings() -> Settings:
     """Get cached settings"""
     return Settings()
+
+
+# Create global settings instance
+settings = get_settings()

backend/app/core/crud.py

@@ -0,0 +1,178 @@
+"""
+Database CRUD Operations
+数据库通用 CRUD 操作
+"""
+from typing import Any, Generic, List, Optional, Type, TypeVar
+from uuid import UUID
+from sqlalchemy import select, func, and_
+from sqlalchemy.ext.asyncio import AsyncSession
+from sqlalchemy.orm import selectinload, joinedload
+
+from app.core.exceptions import NotFoundException, DuplicateException
+from app.core.logging import LoggerMixin
+
+ModelType = TypeVar("ModelType")
+
+
+class CRUDBase(Generic[ModelType], LoggerMixin):
+    """Base CRUD class with common operations"""
+
+    def __init__(self, model: Type[ModelType]):
+        """Initialize CRUD with model"""
+        self.model = model
+
+    async def get(
+        self,
+        db: AsyncSession,
+        id: UUID,
+        load_relations: Optional[List[str]] = None
+    ) -> Optional[ModelType]:
+        """Get single record by ID"""
+        query = select(self.model).where(self.model.id == id)
+
+        # Load relationships if specified
+        if load_relations:
+            for relation in load_relations:
+                if hasattr(self.model, relation):
+                    query = query.options(selectinload(getattr(self.model, relation)))
+
+        result = await db.execute(query)
+        return result.scalar_one_or_none()
+
+    async def get_or_raise(
+        self,
+        db: AsyncSession,
+        id: UUID,
+        resource_name: str = "Resource",
+        load_relations: Optional[List[str]] = None
+    ) -> ModelType:
+        """Get single record by ID or raise NotFoundException"""
+        obj = await self.get(db, id, load_relations)
+        if not obj:
+            raise NotFoundException(resource_name, id)
+        return obj
+
+    async def get_multi(
+        self,
+        db: AsyncSession,
+        skip: int = 0,
+        limit: int = 20,
+        load_relations: Optional[List[str]] = None,
+        filters: Optional[dict] = None,
+        order_by: Optional[str] = "created_at",
+        descending: bool = True
+    ) -> tuple[List[ModelType], int]:
+        """Get multiple records with pagination"""
+        query = select(self.model)
+        count_query = select(func.count()).select_from(self.model)
+
+        # Apply filters
+        if filters:
+            conditions = []
+            for key, value in filters.items():
+                if hasattr(self.model, key):
+                    conditions.append(getattr(self.model, key) == value)
+            if conditions:
+                query = query.where(and_(*conditions))
+                count_query = count_query.where(and_(*conditions))
+
+        # Load relationships if specified
+        if load_relations:
+            for relation in load_relations:
+                if hasattr(self.model, relation):
+                    query = query.options(selectinload(getattr(self.model, relation)))
+
+        # Count total
+        total_result = await db.execute(count_query)
+        total = total_result.scalar() or 0
+
+        # Apply ordering
+        if order_by and hasattr(self.model, order_by):
+            order_column = getattr(self.model, order_by)
+            if descending:
+                query = query.order_by(order_column.desc())
+            else:
+                query = query.order_by(order_column.asc())
+
+        # Apply pagination
+        query = query.offset(skip).limit(limit)
+
+        result = await db.execute(query)
+        items = result.scalars().all()
+
+        return list(items), total
+
+    async def create(
+        self,
+        db: AsyncSession,
+        obj_in: Any,
+        commit: bool = True
+    ) -> ModelType:
+        """Create new record"""
+        obj_data = obj_in.model_dump() if hasattr(obj_in, 'model_dump') else obj_in.dict()
+        db_obj = self.model(**obj_data)
+        db.add(db_obj)
+
+        if commit:
+            await db.commit()
+            await db.refresh(db_obj)
+            self.log.debug(f"Created {self.model.__name__}: {db_obj.id}")
+
+        return db_obj
+
+    async def update(
+        self,
+        db: AsyncSession,
+        db_obj: ModelType,
+        obj_in: Any,
+        commit: bool = True
+    ) -> ModelType:
+        """Update existing record"""
+        if hasattr(obj_in, 'model_dump'):
+            obj_data = obj_in.model_dump(exclude_unset=True)
+        elif hasattr(obj_in, 'dict'):
+            obj_data = obj_in.dict(exclude_unset=True)
+        else:
+            obj_data = obj_in
+
+        for field, value in obj_data.items():
+            if hasattr(db_obj, field):
+                setattr(db_obj, field, value)
+
+        if commit:
+            await db.commit()
+            await db.refresh(db_obj)
+            self.log.debug(f"Updated {self.model.__name__}: {db_obj.id}")
+
+        return db_obj
+
+    async def delete(
+        self,
+        db: AsyncSession,
+        id: UUID,
+        commit: bool = True
+    ) -> bool:
+        """Delete record by ID"""
+        obj = await self.get(db, id)
+        if obj:
+            await db.delete(obj)
+            if commit:
+                await db.commit()
+            self.log.debug(f"Deleted {self.model.__name__}: {id}")
+            return True
+        return False
+
+    async def exists(
+        self,
+        db: AsyncSession,
+        filters: dict
+    ) -> bool:
+        """Check if record exists"""
+        query = select(func.count()).select_from(self.model)
+        for key, value in filters.items():
+            if hasattr(self.model, key):
+                query = query.where(getattr(self.model, key) == value)
+
+        result = await db.execute(query)
+        count = result.scalar() or 0
+        return count > 0

backend/app/core/database.py

@@ -2,25 +2,32 @@
 Database Configuration and Session Management
 支持 SQLite 和 PostgreSQL
 """
-
+import logging
+from contextlib import asynccontextmanager
+from typing import AsyncGenerator
 from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine, async_sessionmaker
 from sqlalchemy.orm import DeclarativeBase
-from sqlalchemy import create_engine
+from sqlalchemy import create_engine, event, inspect, text
+from sqlalchemy.pool import NullPool
+
 from app.core.config import get_settings
 
+logger = logging.getLogger(__name__)
 settings = get_settings()
 
 
 def get_engine_config():
     """根据数据库类型返回引擎配置"""
     if settings.DATABASE_URL.startswith("sqlite"):
-        return {"echo": settings.DEBUG}
+        return {"echo": settings.DEBUG, "poolclass": NullPool}
     else:
         return {
             "echo": settings.DEBUG,
             "pool_pre_ping": True,
             "pool_size": 10,
             "max_overflow": 20,
+            "pool_recycle": 3600,
+            "pool_timeout": 30,
         }
 
 
@@ -30,14 +37,14 @@ async_engine = create_async_engine(
     **get_engine_config()
 )
 
-# Sync engine for migrations
+# Sync engine for migrations (use NullPool for SQLite)
 sync_engine = create_engine(
     settings.DATABASE_URL_SYNC,
     echo=settings.DEBUG,
     pool_pre_ping=True,
+    poolclass=NullPool if settings.DATABASE_URL_SYNC.startswith("sqlite") else None,
 )
 
-
 # Async session factory
 AsyncSessionLocal = async_sessionmaker(
     async_engine,
@@ -55,8 +62,50 @@ class Base(DeclarativeBase):
 
 async def init_db():
     """Initialize database tables"""
+    logger.info("Initializing database...")
     async with async_engine.begin() as conn:
         await conn.run_sync(Base.metadata.create_all)
+        await conn.run_sync(_ensure_legacy_columns)
+    logger.info("Database initialized successfully")
+
+
+def _ensure_legacy_columns(sync_conn):
+    """Patch legacy tables with newly introduced columns."""
+    inspector = inspect(sync_conn)
+    if "model_configs" not in inspector.get_table_names():
+        return
+
+    columns = {column["name"] for column in inspector.get_columns("model_configs")}
+    if "model_type" in columns:
+        return
+
+    logger.info("Adding missing model_type column to model_configs table")
+    dialect = sync_conn.dialect.name
+    if dialect == "postgresql":
+        sync_conn.execute(text("ALTER TABLE model_configs ADD COLUMN model_type VARCHAR(50) NOT NULL DEFAULT 'chat'"))
+    else:
+        sync_conn.execute(text("ALTER TABLE model_configs ADD COLUMN model_type VARCHAR(50) NOT NULL DEFAULT 'chat'"))
+
+
+async def close_db():
+    """Close database connections"""
+    logger.info("Closing database connections...")
+    await async_engine.dispose()
+    logger.info("Database connections closed")
+
+
+@asynccontextmanager
+async def get_db_session() -> AsyncGenerator[AsyncSession, None]:
+    """Context manager for database sessions with automatic cleanup"""
+    session = AsyncSessionLocal()
+    try:
+        yield session
+    except Exception as e:
+        logger.error(f"Database session error: {str(e)}")
+        await session.rollback()
+        raise
+    finally:
+        await session.close()
 
 
 async def get_db() -> AsyncSession:
@@ -64,5 +113,14 @@ async def get_db() -> AsyncSession:
     async with AsyncSessionLocal() as session:
         try:
             yield session
+        except Exception as e:
+            logger.error(f"Database error in dependency: {str(e)}")
+            await session.rollback()
+            raise
         finally:
             await session.close()
+
+
+# Import all models to register them with Base.metadata
+# This ensures all models are loaded before create_all is called
+from app.models.models import *  # noqa: F401, F403, E402

backend/app/core/exceptions.py

@@ -0,0 +1,119 @@
+"""
+Custom Exceptions
+自定义异常类
+"""
+from typing import Any, Optional
+
+
+class AppException(Exception):
+    """基础应用异常"""
+
+    def __init__(
+        self,
+        message: str,
+        code: str = "INTERNAL_ERROR",
+        status_code: int = 500,
+        details: Optional[dict] = None
+    ):
+        self.message = message
+        self.code = code
+        self.status_code = status_code
+        self.details = details
+        super().__init__(self.message)
+
+
+class NotFoundException(AppException):
+    """资源未找到异常"""
+
+    def __init__(self, resource: str, resource_id: Any = None):
+        message = f"{resource} not found"
+        if resource_id:
+            message = f"{resource} with id '{resource_id}' not found"
+        super().__init__(
+            message=message,
+            code="NOT_FOUND",
+            status_code=404
+        )
+
+
+class ValidationException(AppException):
+    """验证异常"""
+
+    def __init__(self, message: str, field: str = None, details: dict = None):
+        super().__init__(
+            message=message,
+            code="VALIDATION_ERROR",
+            status_code=422,
+            details={"field": field, **(details or {})}
+        )
+
+
+class DuplicateException(AppException):
+    """重复资源异常"""
+
+    def __init__(self, resource: str, field: str = None):
+        message = f"{resource} already exists"
+        if field:
+            message = f"{resource} with {field} already exists"
+        super().__init__(
+            message=message,
+            code="DUPLICATE",
+            status_code=409
+        )
+
+
+class UnauthorizedException(AppException):
+    """未授权异常"""
+
+    def __init__(self, message: str = "Unauthorized"):
+        super().__init__(
+            message=message,
+            code="UNAUTHORIZED",
+            status_code=401
+        )
+
+
+class ForbiddenException(AppException):
+    """禁止访问异常"""
+
+    def __init__(self, message: str = "Forbidden"):
+        super().__init__(
+            message=message,
+            code="FORBIDDEN",
+            status_code=403
+        )
+
+
+class RateLimitException(AppException):
+    """速率限制异常"""
+
+    def __init__(self, message: str = "Rate limit exceeded"):
+        super().__init__(
+            message=message,
+            code="RATE_LIMIT",
+            status_code=429
+        )
+
+
+class FileProcessingException(AppException):
+    """文件处理异常"""
+
+    def __init__(self, message: str, file_name: str = None):
+        details = {"file_name": file_name} if file_name else None
+        super().__init__(
+            message=message,
+            code="FILE_PROCESSING_ERROR",
+            status_code=422,
+            details=details
+        )
+
+
+class DatabaseException(AppException):
+    """数据库异常"""
+
+    def __init__(self, message: str = "Database operation failed"):
+        super().__init__(
+            message=message,
+            code="DATABASE_ERROR",
+            status_code=500
+        )

backend/app/core/logging.py

@@ -0,0 +1,139 @@
+"""
+Logging Configuration
+日志配置
+"""
+import logging
+import sys
+from datetime import datetime
+from typing import Any
+from logging.handlers import RotatingFileHandler, TimedRotatingFileHandler
+from pathlib import Path
+from app.core.config import get_settings
+
+settings = get_settings()
+
+# Log directory
+LOG_DIR = Path("./logs")
+LOG_DIR.mkdir(exist_ok=True)
+
+# 日期格式
+LOG_DATE = datetime.now().strftime("%Y-%m-%d")
+
+# 当天的日志目录
+CURRENT_LOG_DIR = LOG_DIR / LOG_DATE
+CURRENT_LOG_DIR.mkdir(exist_ok=True)
+
+
+def get_log_path(filename: str) -> Path:
+    """获取当天的日志文件路径"""
+    return CURRENT_LOG_DIR / filename
+
+
+def setup_logging(name: str = "yg_dataset") -> logging.Logger:
+    """Setup application logging"""
+    logger = logging.getLogger(name)
+    logger.setLevel(logging.DEBUG if settings.DEBUG else logging.INFO)
+
+    # Avoid duplicate handlers
+    if logger.handlers:
+        return logger
+
+    # Console handler
+    console_handler = logging.StreamHandler(sys.stdout)
+    console_handler.setLevel(logging.DEBUG if settings.DEBUG else logging.INFO)
+    console_formatter = logging.Formatter(
+        fmt="%(asctime)s | %(levelname)-8s | %(name)s:%(funcName)s:%(lineno)d | %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S"
+    )
+    console_handler.setFormatter(console_formatter)
+    logger.addHandler(console_handler)
+
+    # Main log file handler - app.log
+    main_file_handler = TimedRotatingFileHandler(
+        get_log_path("app.log"),
+        when="midnight",
+        interval=1,
+        backupCount=30,
+        encoding="utf-8"
+    )
+    main_file_handler.setLevel(logging.INFO)
+    main_file_formatter = logging.Formatter(
+        fmt="%(asctime)s | %(levelname)-8s | %(name)s:%(funcName)s:%(lineno)d | %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S"
+    )
+    main_file_handler.setFormatter(main_file_formatter)
+    logger.addHandler(main_file_handler)
+
+    return logger
+
+
+# Create default logger
+logger = setup_logging()
+
+
+# ============== Success Logger ==============
+def get_success_logger() -> logging.Logger:
+    """获取成功日志记录器"""
+    success_logger = logging.getLogger("yg_dataset.success")
+    if not success_logger.handlers:
+        handler = RotatingFileHandler(
+            get_log_path("success.log"),
+            maxBytes=10 * 1024 * 1024,
+            backupCount=30,
+            encoding="utf-8"
+        )
+        handler.setLevel(logging.INFO)
+        formatter = logging.Formatter(
+            fmt="%(asctime)s | %(message)s",
+            datefmt="%Y-%m-%d %H:%M:%S"
+        )
+        handler.setFormatter(formatter)
+        success_logger.addHandler(handler)
+        success_logger.setLevel(logging.INFO)
+    return success_logger
+
+
+# ============== Failure Logger ==============
+def get_failure_logger() -> logging.Logger:
+    """获取失败日志记录器"""
+    failure_logger = logging.getLogger("yg_dataset.failure")
+    if not failure_logger.handlers:
+        handler = RotatingFileHandler(
+            get_log_path("failure.log"),
+            maxBytes=10 * 1024 * 1024,
+            backupCount=30,
+            encoding="utf-8"
+        )
+        handler.setLevel(logging.WARNING)
+        formatter = logging.Formatter(
+            fmt="%(asctime)s | %(levelname)-8s | %(name)s:%(funcName)s:%(lineno)d | %(message)s",
+            datefmt="%Y-%m-%d %H:%M:%S"
+        )
+        handler.setFormatter(formatter)
+        failure_logger.addHandler(handler)
+        failure_logger.setLevel(logging.WARNING)
+    return failure_logger
+
+
+# ============== Convenience functions ==============
+def log_success(message: str, **kwargs):
+    """记录成功日志"""
+    extra_info = " | ".join([f"{k}={v}" for k, v in kwargs.items()]) if kwargs else ""
+    full_message = f"{message} | {extra_info}" if extra_info else message
+    get_success_logger().info(full_message)
+
+
+def log_failure(message: str, **kwargs):
+    """记录失败日志"""
+    extra_info = " | ".join([f"{k}={v}" for k, v in kwargs.items()]) if kwargs else ""
+    full_message = f"{message} | {extra_info}" if extra_info else message
+    get_failure_logger().warning(full_message)
+
+
+class LoggerMixin:
+    """Mixin to add logging capability to classes"""
+
+    @property
+    def log(self) -> logging.Logger:
+        """Get logger for this class"""
+        return logging.getLogger(self.__class__.__module__ + "." + self.__class__.__name__)

backend/app/main.py

@@ -3,23 +3,74 @@ YG-Dataset Backend Application
 FastAPI-based API server for dataset generation platform
 """
 
+import logging
+import time
+import uuid
 from contextlib import asynccontextmanager
-from fastapi import FastAPI
+from fastapi import FastAPI, Request
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import JSONResponse
+from fastapi.exceptions import RequestValidationError
+from starlette.middleware.base import BaseHTTPMiddleware
+from sqlalchemy.exc import SQLAlchemyError
 
 from app.api.v1 import api_router
+from app.api.response import ApiResponse
 from app.core.config import settings
-from app.core.database import init_db
+from app.core.database import init_db, close_db
+from app.core.exceptions import AppException
+from app.core.logging import logger
+
+# Import all models to register them with Base.metadata
+from app.models.models import *  # noqa: F401, F403
+
+
+class RequestIDMiddleware(BaseHTTPMiddleware):
+    """Middleware to add request ID to each request"""
+
+    async def dispatch(self, request: Request, call_next):
+        request_id = str(uuid.uuid4())
+        request.state.request_id = request_id
+
+        # Add request ID to response headers
+        response = await call_next(request)
+        response.headers["X-Request-ID"] = request_id
+
+        return response
+
+
+class TimingMiddleware(BaseHTTPMiddleware):
+    """Middleware to measure request processing time"""
+
+    async def dispatch(self, request: Request, call_next):
+        start_time = time.time()
+
+        # Log request
+        logger.info(f"→ {request.method} {request.url.path}")
+
+        response = await call_next(request)
+
+        process_time = time.time() - start_time
+        response.headers["X-Process-Time"] = str(process_time)
+
+        # Log response
+        logger.info(f"← {request.method} {request.url.path} | Status: {response.status_code} | Time: {process_time:.3f}s")
+
+        return response
 
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     """Application lifespan events"""
     # Startup
+    logger.info("Starting YG-Dataset application...")
     await init_db()
+    logger.info("Database initialized successfully")
     yield
     # Shutdown
-    pass
+    logger.info("Shutting down YG-Dataset application...")
+    await close_db()
+    logger.info("Database connections closed")
 
 
 app = FastAPI(
@@ -29,15 +80,83 @@ app = FastAPI(
     lifespan=lifespan,
 )
 
-# CORS
+# Add custom middleware (order matters: last added = first executed)
+app.add_middleware(TimingMiddleware)
+app.add_middleware(RequestIDMiddleware)
+
+# CORS - Configure properly for production
+# For development, you can use ["*"] but for production, specify exact origins
+ALLOWED_ORIGINS = ["*"]
+
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["*"],
+    allow_origins=ALLOWED_ORIGINS,
     allow_credentials=True,
-    allow_methods=["*"],
+    allow_methods=["GET", "POST", "PUT", "DELETE", "OPTIONS"],
     allow_headers=["*"],
 )
 
+
+# Exception handlers
+@app.exception_handler(AppException)
+async def app_exception_handler(request: Request, exc: AppException):
+    """Handle custom application exceptions"""
+    logger.warning(f"App exception: {exc.message} | Code: {exc.code}")
+    return JSONResponse(
+        status_code=exc.status_code,
+        content=ApiResponse.fail(
+            message=exc.message,
+            error={"code": exc.code, "details": exc.details}
+        ).model_dump(mode='json')
+    )
+
+
+@app.exception_handler(RequestValidationError)
+async def validation_exception_handler(request: Request, exc: RequestValidationError):
+    """Handle validation exceptions"""
+    errors = []
+    for error in exc.errors():
+        errors.append({
+            "field": ".".join(str(loc) for loc in error["loc"]),
+            "message": error["msg"],
+            "type": error["type"]
+        })
+    logger.warning(f"Validation error: {errors}")
+    return JSONResponse(
+        status_code=422,
+        content=ApiResponse.fail(
+            message="Validation error",
+            error={"code": "VALIDATION_ERROR", "details": {"errors": errors}}
+        ).model_dump(mode='json')
+    )
+
+
+@app.exception_handler(SQLAlchemyError)
+async def database_exception_handler(request: Request, exc: SQLAlchemyError):
+    """Handle database exceptions"""
+    logger.error(f"Database error: {str(exc)}", exc_info=True)
+    return JSONResponse(
+        status_code=500,
+        content=ApiResponse.fail(
+            message="Database operation failed",
+            error={"code": "DATABASE_ERROR"}
+        ).model_dump(mode='json')
+    )
+
+
+@app.exception_handler(Exception)
+async def general_exception_handler(request: Request, exc: Exception):
+    """Handle unhandled exceptions"""
+    logger.error(f"Unhandled exception: {str(exc)}", exc_info=True)
+    return JSONResponse(
+        status_code=500,
+        content=ApiResponse.fail(
+            message="Internal server error",
+            error={"code": "INTERNAL_ERROR"}
+        ).model_dump(mode='json')
+    )
+
+
 # Include API routes
 app.include_router(api_router, prefix="/api/v1")
 
@@ -45,7 +164,10 @@ app.include_router(api_router, prefix="/api/v1")
 @app.get("/health")
 async def health_check():
     """Health check endpoint"""
-    return {"status": "healthy", "version": "1.0.0"}
+    return ApiResponse.ok(
+        data={"status": "healthy", "version": "1.0.0"},
+        message="Service is running"
+    )
 
 
 if __name__ == "__main__":

backend/app/models/models.py

@@ -14,6 +14,7 @@ class Project(Base, UUIDMixin, TimestampMixin):
 
     name = Column(String(255), nullable=False)
     description = Column(Text)
+    type = Column(String(50), default="qa")  # qa, table, database
 
     # Relationships
     files = relationship("File", back_populates="project", cascade="all, delete-orphan")
@@ -51,7 +52,7 @@ class Chunk(Base, UUIDMixin, TimestampMixin):
     content = Column(Text, nullable=False)
     summary = Column(Text)
     word_count = Column(Integer)
-    metadata = Column(JSON)  # store additional info like headings, page numbers
+    extra_data = Column(JSON)  # store additional info like headings, page numbers
 
     # Relationships
     project = relationship("Project", back_populates="chunks")
@@ -112,7 +113,7 @@ class Dataset(Base, UUIDMixin, TimestampMixin):
     name = Column(String(255), nullable=False)
     description = Column(Text)
     dataset_type = Column(String(50))  # qa, conversation, instruction
-    metadata = Column(JSON)
+    extra_data = Column(JSON)
 
     # Relationships
     project = relationship("Project", back_populates="datasets")
@@ -125,7 +126,7 @@ class EvalDataset(Base, UUIDMixin, TimestampMixin):
     project_id = Column(UUID(as_uuid=True), ForeignKey("projects.id", ondelete="CASCADE"), nullable=False)
     name = Column(String(255), nullable=False)
     question_type = Column(String(50))  # mixed, fact, reasoning
-    metadata = Column(JSON)
+    extra_data = Column(JSON)
 
     # Relationships
     project = relationship("Project", back_populates="eval_datasets")
@@ -135,12 +136,14 @@ class ModelConfig(Base, UUIDMixin, TimestampMixin):
     """Model configuration for LLM providers"""
     __tablename__ = "model_configs"
 
-    project_id = Column(UUID(as_uuid=True), ForeignKey("projects.id", ondelete="CASCADE"), nullable=False)
-    provider = Column(String(50), nullable=False)  # openai, anthropic, ollama, custom
+    project_id = Column(UUID(as_uuid=True), ForeignKey("projects.id", ondelete="CASCADE"), nullable=True)
+    provider = Column(String(50), nullable=False)  # minimax, glm, openai, ali
+    model_type = Column(String(50), nullable=False, default="chat")  # chat, vlm, embedding, rerank
     model_name = Column(String(100))
     api_key = Column(String(500))
     api_base = Column(String(500))
     is_default = Column(String(10), default="false")
+    connection_status = Column(String(20), default="untested")  # untested, connected, disconnected
 
     # Relationships
     project = relationship("Project", back_populates="model_configs")

backend/app/schemas/__init__.py

@@ -1,3 +1,101 @@
 """
 Pydantic Schemas
 """
+from app.schemas.base import (
+    TimestampMixin,
+    UUIDMixin,
+)
+
+from app.schemas.project import (
+    ProjectBase,
+    ProjectCreate,
+    ProjectUpdate,
+    ProjectResponse,
+)
+
+from app.schemas.file import (
+    FileBase,
+    FileCreate,
+    FileUpdate,
+    FileResponse,
+)
+
+from app.schemas.chunk import (
+    ChunkBase,
+    ChunkCreate,
+    ChunkUpdate,
+    ChunkResponse,
+)
+
+from app.schemas.question import (
+    QuestionBase,
+    QuestionCreate,
+    QuestionUpdate,
+    QuestionResponse,
+)
+
+from app.schemas.dataset import (
+    DatasetBase,
+    DatasetCreate,
+    DatasetUpdate,
+    DatasetResponse,
+)
+
+from app.schemas.eval import (
+    EvalDatasetBase,
+    EvalDatasetCreate,
+    EvalDatasetUpdate,
+    EvalDatasetResponse,
+    TaskBase,
+    TaskResponse,
+)
+
+from app.schemas.model import (
+    ModelBase,
+    ModelCreate,
+    ModelUpdate,
+    ModelResponse,
+)
+
+__all__ = [
+    # Base
+    "TimestampMixin",
+    "UUIDMixin",
+    # Project
+    "ProjectBase",
+    "ProjectCreate",
+    "ProjectUpdate",
+    "ProjectResponse",
+    # File
+    "FileBase",
+    "FileCreate",
+    "FileUpdate",
+    "FileResponse",
+    # Chunk
+    "ChunkBase",
+    "ChunkCreate",
+    "ChunkUpdate",
+    "ChunkResponse",
+    # Question
+    "QuestionBase",
+    "QuestionCreate",
+    "QuestionUpdate",
+    "QuestionResponse",
+    # Dataset
+    "DatasetBase",
+    "DatasetCreate",
+    "DatasetUpdate",
+    "DatasetResponse",
+    # Eval
+    "EvalDatasetBase",
+    "EvalDatasetCreate",
+    "EvalDatasetUpdate",
+    "EvalDatasetResponse",
+    "TaskBase",
+    "TaskResponse",
+    # Model
+    "ModelBase",
+    "ModelCreate",
+    "ModelUpdate",
+    "ModelResponse",
+]

backend/app/schemas/base.py

@@ -4,7 +4,7 @@ Base Pydantic schemas
 from datetime import datetime
 from typing import Optional, Any
 from uuid import UUID
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, ConfigDict, Field
 
 
 class TimestampMixin(BaseModel):
@@ -18,153 +18,3 @@ class UUIDMixin(BaseModel):
     model_config = ConfigDict(from_attributes=True)
 
     id: UUID
-
-
-class ProjectBase(BaseModel):
-    """Base project schema"""
-    name: str
-    description: Optional[str] = None
-
-
-class ProjectCreate(ProjectBase):
-    """Project create schema"""
-    pass
-
-
-class ProjectUpdate(ProjectBase):
-    """Project update schema"""
-    pass
-
-
-class ProjectResponse(ProjectBase, UUIDMixin, TimestampMixin):
-    """Project response schema"""
-    pass
-
-
-class FileBase(BaseModel):
-    """Base file schema"""
-    filename: str
-    file_type: str
-    size: Optional[int] = None
-
-
-class FileResponse(FileBase, UUIDMixin, TimestampMixin):
-    """File response schema"""
-    status: str
-
-
-class ChunkBase(BaseModel):
-    """Base chunk schema"""
-    name: Optional[str] = None
-    content: str
-    summary: Optional[str] = None
-    word_count: Optional[int] = None
-
-
-class ChunkCreate(ChunkBase):
-    """Chunk create schema"""
-    file_id: Optional[UUID] = None
-
-
-class ChunkResponse(ChunkBase, UUIDMixin, TimestampMixin):
-    """Chunk response schema"""
-    pass
-
-
-class QuestionBase(BaseModel):
-    """Base question schema"""
-    content: str
-    answer: Optional[str] = None
-    question_type: Optional[str] = None
-
-
-class QuestionCreate(QuestionBase):
-    """Question create schema"""
-    chunk_id: Optional[UUID] = None
-
-
-class QuestionResponse(QuestionBase, UUIDMixin, TimestampMixin):
-    """Question response schema"""
-    source: str
-
-
-class DatasetBase(BaseModel):
-    """Base dataset schema"""
-    name: str
-    description: Optional[str] = None
-    dataset_type: Optional[str] = None
-
-
-class DatasetCreate(DatasetBase):
-    """Dataset create schema"""
-    pass
-
-
-class DatasetResponse(DatasetBase, UUIDMixin, TimestampMixin):
-    """Dataset response schema"""
-    question_count: Optional[int] = None
-
-
-class EvalDatasetBase(BaseModel):
-    """Base eval dataset schema"""
-    name: str
-    question_type: Optional[str] = None
-
-
-class EvalDatasetCreate(EvalDatasetBase):
-    """Eval dataset create schema"""
-    pass
-
-
-class EvalDatasetResponse(EvalDatasetBase, UUIDMixin, TimestampMixin):
-    """Eval dataset response schema"""
-    pass
-
-
-class TagBase(BaseModel):
-    """Base tag schema"""
-    label: str
-    parent_id: Optional[UUID] = None
-    color: Optional[str] = None
-
-
-class TagCreate(TagBase):
-    """Tag create schema"""
-    pass
-
-
-class TagResponse(TagBase, UUIDMixin, TimestampMixin):
-    """Tag response schema"""
-    pass
-
-
-class ModelConfigBase(BaseModel):
-    """Base model config schema"""
-    provider: str
-    model_name: Optional[str] = None
-    api_key: Optional[str] = None
-    api_base: Optional[str] = None
-    is_default: Optional[str] = "false"
-
-
-class ModelConfigCreate(ModelConfigBase):
-    """Model config create schema"""
-    pass
-
-
-class ModelConfigResponse(ModelConfigBase, UUIDMixin, TimestampMixin):
-    """Model config response schema"""
-    pass
-
-
-class TaskBase(BaseModel):
-    """Base task schema"""
-    task_type: str
-    status: Optional[str] = "pending"
-    progress: Optional[int] = 0
-
-
-class TaskResponse(TaskBase, UUIDMixin, TimestampMixin):
-    """Task response schema"""
-    result: Optional[Any] = None
-    error: Optional[str] = None

backend/app/schemas/chunk.py

@@ -0,0 +1,46 @@
+"""
+Chunk Schemas
+"""
+from datetime import datetime
+from typing import Optional, Any
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class ChunkBase(BaseModel):
+    """Base chunk schema"""
+    name: Optional[str] = Field(None, max_length=255)
+    content: str = Field(..., min_length=1)
+    summary: Optional[str] = None
+    word_count: Optional[int] = None
+    extra_data: Optional[dict] = None
+
+
+class ChunkCreate(ChunkBase):
+    """Chunk create schema"""
+    project_id: Optional[UUID] = None
+    file_id: Optional[UUID] = None
+
+
+class ChunkUpdate(BaseModel):
+    """Chunk update schema"""
+    name: Optional[str] = Field(None, max_length=255)
+    content: Optional[str] = Field(None, min_length=1)
+    summary: Optional[str] = None
+    extra_data: Optional[dict] = None
+
+
+class ChunkResponse(ChunkBase):
+    """Chunk response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    project_id: UUID
+    file_id: Optional[UUID]
+    created_at: datetime
+    updated_at: datetime
+
+
+# Alias for CRUD
+ChunkCreateSchema = ChunkCreate
+ChunkUpdateSchema = ChunkUpdate

backend/app/schemas/dataset.py

@@ -0,0 +1,43 @@
+"""
+Dataset Schemas
+"""
+from datetime import datetime
+from typing import Optional, Any
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class DatasetBase(BaseModel):
+    """Base dataset schema"""
+    name: str = Field(..., min_length=1, max_length=255)
+    description: Optional[str] = Field(None, max_length=2000)
+    dataset_type: Optional[str] = Field(None, max_length=50)
+    extra_data: Optional[dict] = None
+
+
+class DatasetCreate(DatasetBase):
+    """Dataset create schema"""
+    pass
+
+
+class DatasetUpdate(BaseModel):
+    """Dataset update schema"""
+    name: Optional[str] = Field(None, min_length=1, max_length=255)
+    description: Optional[str] = Field(None, max_length=2000)
+    dataset_type: Optional[str] = Field(None, max_length=50)
+    extra_data: Optional[dict] = None
+
+
+class DatasetResponse(DatasetBase):
+    """Dataset response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    project_id: UUID
+    created_at: datetime
+    updated_at: datetime
+
+
+# Alias for CRUD
+DatasetCreateSchema = DatasetCreate
+DatasetUpdateSchema = DatasetUpdate

backend/app/schemas/eval.py

@@ -0,0 +1,60 @@
+"""
+Evaluation Dataset Schemas
+"""
+from datetime import datetime
+from typing import Optional, Any
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class EvalDatasetBase(BaseModel):
+    """Base eval dataset schema"""
+    name: str = Field(..., min_length=1, max_length=255)
+    question_type: Optional[str] = Field("mixed", max_length=50)
+    extra_data: Optional[dict] = None
+
+
+class EvalDatasetCreate(EvalDatasetBase):
+    """Eval dataset create schema"""
+    pass
+
+
+class EvalDatasetUpdate(BaseModel):
+    """Eval dataset update schema"""
+    name: Optional[str] = Field(None, min_length=1, max_length=255)
+    question_type: Optional[str] = Field(None, max_length=50)
+    extra_data: Optional[dict] = None
+
+
+class EvalDatasetResponse(EvalDatasetBase):
+    """Eval dataset response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    project_id: UUID
+    created_at: datetime
+    updated_at: datetime
+
+
+class TaskBase(BaseModel):
+    """Base task schema"""
+    task_type: str = Field(..., max_length=50)
+    status: Optional[str] = "pending"
+    progress: Optional[int] = Field(0, ge=0, le=100)
+    result: Optional[Any] = None
+    error: Optional[str] = None
+
+
+class TaskResponse(TaskBase):
+    """Task response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    project_id: Optional[UUID]
+    created_at: datetime
+    updated_at: datetime
+
+
+# Alias for CRUD
+EvalDatasetCreateSchema = EvalDatasetCreate
+EvalDatasetUpdateSchema = EvalDatasetUpdate

backend/app/schemas/file.py

@@ -0,0 +1,43 @@
+"""
+File Schemas
+"""
+from datetime import datetime
+from typing import Optional
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class FileBase(BaseModel):
+    """Base file schema"""
+    filename: str = Field(..., min_length=1, max_length=255)
+    file_type: str = Field(..., max_length=50)
+    size: Optional[int] = None
+
+
+class FileCreate(FileBase):
+    """File create schema"""
+    project_id: UUID
+    file_path: Optional[str] = None
+    status: str = "pending"
+
+
+class FileUpdate(BaseModel):
+    """File update schema"""
+    status: Optional[str] = None
+
+
+class FileResponse(FileBase):
+    """File response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    project_id: UUID
+    file_path: Optional[str]
+    status: str
+    created_at: datetime
+    updated_at: datetime
+
+
+# Alias for CRUD
+FileCreateSchema = FileCreate
+FileUpdateSchema = FileUpdate

backend/app/schemas/model.py

@@ -0,0 +1,43 @@
+"""
+Model Schema
+"""
+from pydantic import BaseModel, Field, ConfigDict
+from typing import Optional
+from datetime import datetime
+from uuid import UUID
+
+
+class ModelBase(BaseModel):
+    """Base model schema"""
+    provider: str = Field(..., description="Model provider: minimax, glm, openai, ali")
+    model_type: str = Field(default="chat", description="Model type: chat, vlm, embedding, rerank")
+    model_name: str = Field(..., description="Model name")
+    api_key: Optional[str] = Field(None, description="API key")
+    api_base: Optional[str] = Field(None, description="API base URL")
+    is_default: str = Field(default="false", description="Is default model: true/false")
+
+
+class ModelCreate(ModelBase):
+    """Model creation schema"""
+    pass
+
+
+class ModelUpdate(BaseModel):
+    """Model update schema"""
+    provider: Optional[str] = None
+    model_type: Optional[str] = None
+    model_name: Optional[str] = None
+    api_key: Optional[str] = None
+    api_base: Optional[str] = None
+    is_default: Optional[str] = None
+
+
+class ModelResponse(ModelBase):
+    """Model response schema"""
+    id: UUID
+    created_at: Optional[datetime] = None
+    updated_at: Optional[datetime] = None
+    project_id: Optional[UUID] = None
+    connection_status: Optional[str] = Field(default="untested")
+
+    model_config = ConfigDict(from_attributes=True)

backend/app/schemas/project.py

@@ -0,0 +1,40 @@
+"""
+Project Schemas
+"""
+from datetime import datetime
+from typing import Optional
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class ProjectBase(BaseModel):
+    """Base project schema"""
+    name: str = Field(..., min_length=1, max_length=255)
+    description: Optional[str] = Field(None, max_length=2000)
+    type: str = Field(default="qa")  # qa, table, database
+
+
+class ProjectCreate(ProjectBase):
+    """Project create schema"""
+    pass
+
+
+class ProjectUpdate(BaseModel):
+    """Project update schema"""
+    name: Optional[str] = Field(None, min_length=1, max_length=255)
+    description: Optional[str] = Field(None, max_length=2000)
+    type: Optional[str] = Field(None)
+
+
+class ProjectResponse(ProjectBase):
+    """Project response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    created_at: datetime
+    updated_at: datetime
+
+
+# Alias for CRUD
+ProjectCreateSchema = ProjectCreate
+ProjectUpdateSchema = ProjectUpdate

backend/app/schemas/question.py

@@ -0,0 +1,43 @@
+"""
+Question Schemas
+"""
+from datetime import datetime
+from typing import Optional
+from uuid import UUID
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class QuestionBase(BaseModel):
+    """Base question schema"""
+    content: str = Field(..., min_length=1)
+    answer: Optional[str] = None
+    question_type: Optional[str] = Field(None, max_length=50)
+    source: Optional[str] = "manual"
+
+
+class QuestionCreate(QuestionBase):
+    """Question create schema"""
+    chunk_id: Optional[UUID] = None
+
+
+class QuestionUpdate(BaseModel):
+    """Question update schema"""
+    content: Optional[str] = Field(None, min_length=1)
+    answer: Optional[str] = None
+    question_type: Optional[str] = Field(None, max_length=50)
+
+
+class QuestionResponse(QuestionBase):
+    """Question response schema"""
+    model_config = ConfigDict(from_attributes=True)
+
+    id: UUID
+    project_id: UUID
+    chunk_id: Optional[UUID]
+    created_at: datetime
+    updated_at: datetime
+
+
+# Alias for CRUD
+QuestionCreateSchema = QuestionCreate
+QuestionUpdateSchema = QuestionUpdate

backend/app/services/file_processor/docx_processor.py

@@ -1,8 +1,9 @@
 """
 DOCX Text Extractor
 """
+import asyncio
+from typing import Dict
 from docx import Document
-from typing import Dict, List
 
 
 class DOCXProcessor:
@@ -26,6 +27,12 @@ class DOCXProcessor:
 
         return "\n\n".join(text_parts)
 
+    async def extract_text_async(self, file_path: str) -> str:
+        """Extract all text from DOCX asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_text, file_path
+        )
+
     def extract_with_metadata(self, file_path: str) -> Dict:
         """Extract text with DOCX metadata"""
         doc = Document(file_path)
@@ -46,8 +53,14 @@ class DOCXProcessor:
 
         return result
 
+    async def extract_with_metadata_async(self, file_path: str) -> Dict:
+        """Extract with metadata asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_with_metadata, file_path
+        )
 
-def process_docx(file_path: str) -> str:
+
+async def process_docx(file_path: str) -> str:
     """Process DOCX file and return text"""
     processor = DOCXProcessor()
-    return processor.extract_text(file_path)
+    return await processor.extract_text_async(file_path)

backend/app/services/file_processor/excel_processor.py

@@ -1,8 +1,9 @@
 """
 Excel/CSV Text Extractor
 """
-import pandas as pd
+import asyncio
 from typing import Dict, List
+import pandas as pd
 
 
 class ExcelProcessor:
@@ -13,6 +14,12 @@ class ExcelProcessor:
         df = pd.read_csv(file_path)
         return self._dataframe_to_text(df)
 
+    async def extract_csv_async(self, file_path: str) -> str:
+        """Extract CSV asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_csv, file_path
+        )
+
     def extract_excel(self, file_path: str, sheet_name: str = None) -> str:
         """Extract text from Excel file"""
         if sheet_name:
@@ -27,6 +34,12 @@ class ExcelProcessor:
                 text_parts.append(self._dataframe_to_text(df))
             return "\n\n".join(text_parts)
 
+    async def extract_excel_async(self, file_path: str, sheet_name: str = None) -> str:
+        """Extract Excel asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_excel, file_path, sheet_name
+        )
+
     def _dataframe_to_text(self, df: pd.DataFrame) -> str:
         """Convert DataFrame to readable text"""
         text_parts = []
@@ -48,19 +61,25 @@ class ExcelProcessor:
         sheets = pd.read_excel(file_path, sheet_name=None)
         return {name: self._dataframe_to_text(df) for name, df in sheets.items()}
 
+    async def extract_all_sheets_async(self, file_path: str) -> Dict[str, str]:
+        """Extract all sheets asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_all_sheets, file_path
+        )
+
     def get_sheet_names(self, file_path: str) -> List[str]:
         """Get all sheet names from Excel file"""
         xl = pd.ExcelFile(file_path)
         return xl.sheet_names
 
 
-def process_csv(file_path: str) -> str:
+async def process_csv(file_path: str) -> str:
     """Process CSV file and return text"""
     processor = ExcelProcessor()
-    return processor.extract_csv(file_path)
+    return await processor.extract_csv_async(file_path)
 
 
-def process_excel(file_path: str) -> str:
+async def process_excel(file_path: str) -> str:
     """Process Excel file and return text"""
     processor = ExcelProcessor()
-    return processor.extract_excel(file_path)
+    return await processor.extract_excel_async(file_path)

backend/app/services/file_processor/pdf_processor.py

@@ -1,8 +1,9 @@
 """
 PDF Text Extractor
 """
+import asyncio
+from typing import Dict, List
 import pdfplumber
-from typing import Dict, List, Optional
 
 
 class PDFProcessor:
@@ -20,6 +21,12 @@ class PDFProcessor:
 
         return "\n\n".join(text_parts)
 
+    async def extract_text_async(self, file_path: str) -> str:
+        """Extract all text from PDF asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_text, file_path
+        )
+
     def extract_pages(self, file_path: str) -> List[Dict]:
         """Extract text page by page with metadata"""
         pages = []
@@ -36,6 +43,12 @@ class PDFProcessor:
 
         return pages
 
+    async def extract_pages_async(self, file_path: str) -> List[Dict]:
+        """Extract pages asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_pages, file_path
+        )
+
     def extract_with_metadata(self, file_path: str) -> Dict:
         """Extract text with PDF metadata"""
         result = {
@@ -58,8 +71,14 @@ class PDFProcessor:
 
         return result
 
+    async def extract_with_metadata_async(self, file_path: str) -> Dict:
+        """Extract with metadata asynchronously"""
+        return await asyncio.get_event_loop().run_in_executor(
+            None, self.extract_with_metadata, file_path
+        )
 
-def process_pdf(file_path: str) -> str:
+
+async def process_pdf(file_path: str) -> str:
     """Process PDF file and return text"""
     processor = PDFProcessor()
-    return processor.extract_with_metadata(file_path)["text"]
+    return await processor.extract_with_metadata_async(file_path)

backend/app/services/text_splitter/semantic_embedding.py

@@ -0,0 +1,407 @@
+"""
+Semantic Text Splitter using Online Embedding APIs
+基于在线 Embedding API 的语义分割器
+"""
+import re
+import asyncio
+import httpx
+import numpy as np
+from typing import List, Dict, Optional
+from abc import ABC, abstractmethod
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+
+
+class EmbeddingProvider(ABC):
+    """Embedding API 提供商基类"""
+
+    @abstractmethod
+    async def get_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """获取文本的嵌入向量"""
+        pass
+
+
+class OpenAIEmbedding(EmbeddingProvider):
+    """OpenAI 兼容的 Embedding API"""
+
+    def __init__(self, api_key: str, base_url: str, model: str = "text-embedding-3-small"):
+        self.api_key = api_key
+        self.base_url = base_url.rstrip('/')
+        self.model = model
+
+    async def get_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """调用 OpenAI 兼容的 Embedding API"""
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            headers = {
+                "Authorization": f"Bearer {self.api_key}",
+                "Content-Type": "application/json"
+            }
+
+            # OpenAI 格式
+            payload = {
+                "input": texts,
+                "model": self.model
+            }
+
+            response = await client.post(
+                f"{self.base_url}/embeddings",
+                headers=headers,
+                json=payload
+            )
+            response.raise_for_status()
+            data = response.json()
+
+            # 提取 embeddings
+            return [item["embedding"] for item in data["data"]]
+
+
+class MiniMaxEmbedding(EmbeddingProvider):
+    """MiniMax Embedding API"""
+
+    def __init__(self, api_key: str, base_url: str = "https://api.minimax.chat/v1"):
+        self.api_key = api_key
+        self.base_url = base_url.rstrip('/')
+
+    async def get_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """调用 MiniMax Embedding API"""
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            headers = {
+                "Authorization": f"Bearer {self.api_key}",
+                "Content-Type": "application/json"
+            }
+
+            # MiniMax 格式
+            payload = {
+                "texts": texts,
+                "model": "embo-01"
+            }
+
+            response = await client.post(
+                f"{self.base_url}/text_embeddings",
+                headers=headers,
+                json=payload
+            )
+            response.raise_for_status()
+            data = response.json()
+
+            # MiniMax 返回格式可能不同，需要适配
+            if "data" in data:
+                return [item["embedding"] for item in data["data"]]
+            return []
+
+
+class EmbeddingSplitter:
+    """基于 Embedding 的语义分割器基类"""
+
+    def __init__(
+        self,
+        chunk_size: int = 500,
+        overlap: int = 50,
+        embedding_provider: Optional[EmbeddingProvider] = None,
+        similarity_threshold: float = 0.3,
+        min_chunk_size: int = 100,
+        window_size: int = 3
+    ):
+        self.chunk_size = chunk_size
+        self.overlap = overlap
+        self.embedding_provider = embedding_provider
+        self.similarity_threshold = similarity_threshold
+        self.min_chunk_size = min_chunk_size
+        self.window_size = window_size
+
+    def _tokenize_sentences(self, text: str) -> List[str]:
+        """将文本切分为句子"""
+        paragraphs = re.split(r'\n\s*\n+', text)
+        sentences = []
+        for para in paragraphs:
+            para = para.strip()
+            if not para:
+                continue
+
+            parts = re.split(r'(?<=[。！？；.!?])\s+|(?<=[。！？；])', para)
+            buffer = []
+
+            for part in parts:
+                part = part.strip()
+                if not part:
+                    continue
+
+                # 过短的片段先暂存，尽量与后一句合并，避免 embedding 粒度过碎
+                if len(part) < 8 and buffer:
+                    buffer[-1] = f"{buffer[-1]} {part}".strip()
+                else:
+                    buffer.append(part)
+
+            sentences.extend(buffer)
+
+        return sentences
+
+    def _compute_similarities(self, embeddings: List[List[float]]) -> List[float]:
+        """计算相邻句子的余弦相似度"""
+        similarities = []
+
+        for i in range(len(embeddings) - 1):
+            # 余弦相似度
+            vec1 = np.array(embeddings[i])
+            vec2 = np.array(embeddings[i + 1])
+
+            # 归一化
+            vec1 = vec1 / (np.linalg.norm(vec1) + 1e-8)
+            vec2 = vec2 / (np.linalg.norm(vec2) + 1e-8)
+
+            # 点积 = 余弦相似度（归一化后）
+            sim = np.dot(vec1, vec2)
+            similarities.append(float(sim))
+
+        return similarities
+
+    def _smooth_similarities(self, similarities: List[float]) -> List[float]:
+        """滑动窗口平滑相似度"""
+        if not similarities:
+            return []
+
+        window = max(1, self.window_size)
+        smoothed = []
+
+        for i in range(len(similarities)):
+            start = max(0, i - window)
+            end = min(len(similarities), i + window + 1)
+            window_vals = similarities[start:end]
+            smoothed.append(sum(window_vals) / len(window_vals))
+
+        return smoothed
+
+    def _detect_boundaries(self, similarities: List[float], sentence_lengths: List[int]) -> List[int]:
+        """检测分割点（相似度显著下降的位置）"""
+        if not similarities:
+            return [0]
+
+        smoothed = self._smooth_similarities(similarities)
+        if len(smoothed) <= 1:
+            return [0]
+
+        mean_sim = float(np.mean(smoothed))
+        std_sim = float(np.std(smoothed))
+        dynamic_threshold = max(0.0, min(0.95, mean_sim - 0.5 * std_sim))
+        effective_threshold = max(self.similarity_threshold, dynamic_threshold)
+
+        boundaries = [0]  # 起始点
+        accumulated_chars = 0
+
+        for i, sim in enumerate(smoothed):
+            accumulated_chars += sentence_lengths[i]
+
+            left_sim = smoothed[i - 1] if i > 0 else 1.0
+            right_sim = smoothed[i + 1] if i < len(smoothed) - 1 else 1.0
+            is_local_min = sim <= left_sim and sim <= right_sim
+            has_enough_context = accumulated_chars >= self.min_chunk_size
+            oversize_guard = accumulated_chars >= self.chunk_size
+
+            if (is_local_min and has_enough_context and sim <= effective_threshold) or oversize_guard:
+                boundaries.append(i + 1)
+                accumulated_chars = 0
+
+        boundaries.append(len(sentence_lengths))
+
+        return sorted(list(set(boundaries)))
+
+    def _assemble_chunks(self, sentences: List[str], boundaries: List[int]) -> List[Dict]:
+        """按分割点组装 chunks"""
+        if not sentences:
+            return []
+
+        # 重新计算 boundaries（确保不超过句子数）
+        if not boundaries or boundaries[0] != 0:
+            boundaries = [0] + boundaries
+        if boundaries[-1] != len(sentences):
+            boundaries.append(len(sentences))
+
+        chunks = []
+        for i in range(len(boundaries) - 1):
+            start = boundaries[i]
+            end = boundaries[i + 1]
+            if start >= end:
+                continue
+
+            chunk_text = ' '.join(sentences[start:end]).strip()
+            if not chunk_text:
+                continue
+
+            # 如果 chunk 过大，递归分割
+            if len(chunk_text) > self.chunk_size * 1.5:
+                # 使用更小的窗口再次分割
+                sub_chunks = self._split_large_chunk(sentences[start:end])
+                for j, sub in enumerate(sub_chunks):
+                    chunks.append({
+                        "index": len(chunks),
+                        "content": sub.strip(),
+                        "word_count": len(sub.split()),
+                        "char_count": len(sub)
+                    })
+            else:
+                chunks.append({
+                    "index": len(chunks),
+                    "content": chunk_text.strip(),
+                    "word_count": len(chunk_text.split()),
+                    "char_count": len(chunk_text)
+                })
+
+        # 合并过小的相邻 chunks
+        chunks = self._merge_small_chunks(chunks)
+
+        return chunks
+
+    def _split_large_chunk(self, sentences: List[str]) -> List[str]:
+        """分割过大的 chunk"""
+        # 使用固定长度分割
+        result = []
+        current = ""
+
+        for sent in sentences:
+            if len(current) + len(sent) > self.chunk_size:
+                if current:
+                    result.append(current)
+                current = sent
+            else:
+                current += " " + sent if current else sent
+
+        if current:
+            result.append(current)
+
+        return result
+
+    def _merge_small_chunks(self, chunks: List[Dict]) -> List[Dict]:
+        """合并过小的相邻 chunks"""
+        if len(chunks) <= 1:
+            return chunks
+
+        merged = [chunks[0]]
+
+        for chunk in chunks[1:]:
+            previous = merged[-1]
+            should_merge = (
+                previous["char_count"] < self.min_chunk_size or
+                chunk["char_count"] < self.min_chunk_size
+            )
+
+            if should_merge and previous["char_count"] + chunk["char_count"] <= self.chunk_size * 1.5:
+                previous["content"] += " " + chunk["content"]
+                previous["word_count"] += chunk["word_count"]
+                previous["char_count"] += chunk["char_count"]
+            else:
+                merged.append(chunk)
+
+        for index, chunk in enumerate(merged):
+            chunk["index"] = index
+
+        return merged
+
+    async def split_with_embedding(self, text: str) -> List[Dict]:
+        """使用 Embedding 进行语义分割"""
+        # 1. 句子切分
+        sentences = self._tokenize_sentences(text)
+        if not sentences:
+            return []
+
+        # 过滤纯噪音片段，但保留正常短句
+        sentences = [s for s in sentences if len(s.strip()) >= 4]
+
+        if not sentences:
+            return []
+
+        # 2. 如果只有一个句子，直接返回
+        if len(sentences) == 1:
+            return [{
+                "index": 0,
+                "content": sentences[0],
+                "word_count": len(sentences[0].split()),
+                "char_count": len(sentences[0])
+            }]
+
+        # 3. 调用 Embedding API
+        try:
+            if self.embedding_provider is None:
+                raise ValueError("embedding provider is not configured")
+            embeddings = await self.embedding_provider.get_embeddings(sentences)
+        except Exception as e:
+            # 如果 embedding 失败，降级到规则分割
+            print(f"Embedding failed, falling back to rule-based: {e}")
+            return self._fallback_split(text)
+
+        if len(embeddings) != len(sentences):
+            return self._fallback_split(text)
+
+        # 4. 计算相似度
+        similarities = self._compute_similarities(embeddings)
+
+        # 5. 检测分割点
+        boundaries = self._detect_boundaries(similarities, [len(sentence) for sentence in sentences])
+
+        # 6. 组装 chunks
+        chunks = self._assemble_chunks(sentences, boundaries)
+
+        return chunks
+
+    def _fallback_split(self, text: str) -> List[Dict]:
+        """降级到规则分割"""
+        # 使用 langchain 的 RecursiveCharacterTextSplitter
+        splitter = RecursiveCharacterTextSplitter(
+            chunk_size=self.chunk_size,
+            chunk_overlap=self.overlap,
+            separators=["\n\n", "\n", "。", "！", "？", ". ", "! ", "? "]
+        )
+        chunks = splitter.split_text(text)
+        return [{
+            "index": i,
+            "content": c.strip(),
+            "word_count": len(c.split()),
+            "char_count": len(c)
+        } for i, c in enumerate(chunks)]
+
+
+class SemanticEmbeddingSplitter(EmbeddingSplitter):
+    """基于在线 Embedding 的语义分割器"""
+
+    def __init__(
+        self,
+        chunk_size: int = 500,
+        overlap: int = 50,
+        embedding_provider: Optional[EmbeddingProvider] = None,
+        similarity_threshold: float = 0.3,
+        min_chunk_size: int = 100,
+        window_size: int = 3
+    ):
+        super().__init__(
+            chunk_size=chunk_size,
+            overlap=overlap,
+            embedding_provider=embedding_provider,
+            similarity_threshold=similarity_threshold,
+            min_chunk_size=min_chunk_size,
+            window_size=window_size
+        )
+
+    def split(self, text: str) -> List[Dict]:
+        """同步接口，内部调用异步"""
+        # 由于 split 是同步方法，需要创建新的事件循环
+        try:
+            loop = asyncio.get_event_loop()
+            if loop.is_running():
+                # 如果在异步环境中，创建新任务
+                import concurrent.futures
+                with concurrent.futures.ThreadPoolExecutor() as pool:
+                    future = pool.submit(asyncio.run, self.split_with_embedding(text))
+                    return future.result()
+            else:
+                return loop.run_until_complete(self.split_with_embedding(text))
+        except RuntimeError:
+            # 没有事件循环，直接创建
+            return asyncio.run(self.split_with_embedding(text))
+
+
+def create_embedding_provider(provider: str, api_key: str, base_url: str, model: str = None) -> EmbeddingProvider:
+    """创建 Embedding 提供商"""
+    if provider in ["openai", "compatible", "ali", "glm"]:
+        return OpenAIEmbedding(api_key, base_url, model or "text-embedding-3-small")
+    elif provider == "minimax":
+        return MiniMaxEmbedding(api_key, base_url)
+    else:
+        raise ValueError(f"Unsupported embedding provider: {provider}")

backend/app/services/text_splitter/splitter.py

@@ -3,6 +3,7 @@ Text Splitter
 """
 import re
 from typing import List, Dict, Optional
+from langchain_text_splitters import RecursiveCharacterTextSplitter
 
 
 class TextSplitter:
@@ -18,51 +19,29 @@ class TextSplitter:
 
 
 class RecursiveTextSplitter(TextSplitter):
-    """Recursive character text splitter"""
+    """Recursive character text splitter using langchain"""
 
     def __init__(self, chunk_size: int = 500, overlap: int = 50, separators: List[str] = None):
         super().__init__(chunk_size, overlap)
-        self.separators = separators or ["\n\n", "\n", ". ", " ", ""]
+        self.splitter = RecursiveCharacterTextSplitter(
+            chunk_size=chunk_size,
+            chunk_overlap=overlap,
+            separators=separators or [
+                "\n\n", "\n", ". ", " ", ",", ""
+            ]
+        )
 
     def split(self, text: str) -> List[Dict]:
         """Split text recursively"""
-        chunks = []
-        current_chunk = ""
-        chunk_index = 0
-
-        for separator in self.separators:
-            if separator in text:
-                parts = text.split(separator)
-                for part in parts:
-                    if len(current_chunk) + len(part) > self.chunk_size:
-                        if current_chunk:
-                            chunks.append({
-                                "index": chunk_index,
-                                "content": current_chunk.strip(),
-                                "word_count": len(current_chunk.split())
+        chunks = self.splitter.split_text(text)
+        result = []
+        for i, chunk in enumerate(chunks):
+            result.append({
+                "index": i,
+                "content": chunk.strip(),
+                "word_count": len(chunk.split())
             })
-                            chunk_index += 1
-
-                            # Handle overlap
-                            if self.overlap > 0 and chunks:
-                                overlap_text = " ".join(chunks[-1]["content"].split()[-self.overlap:])
-                                current_chunk = overlap_text + separator + part
-                            else:
-                                current_chunk = part
-                    else:
-                        current_chunk += separator + part if current_chunk else part
-
-                if current_chunk:
-                    chunks.append({
-                        "index": chunk_index,
-                        "content": current_chunk.strip(),
-                        "word_count": len(current_chunk.split())
-                    })
-                break
-            else:
-                continue
-
-        return chunks
+        return result
 
 
 class MarkdownStructureSplitter(TextSplitter):
@@ -236,13 +215,199 @@ class CustomSplitter(TextSplitter):
 
 def get_splitter(method: str, **kwargs) -> TextSplitter:
     """Get text splitter by method name"""
+    # 导入 embedding 分割器
+    from .semantic_embedding import (
+        SemanticEmbeddingSplitter,
+        create_embedding_provider
+    )
+
     splitters = {
         "recursive": RecursiveTextSplitter,
         "markdown_structure": MarkdownStructureSplitter,
         "token": TokenSplitter,
         "code": CodeSplitter,
-        "custom": CustomSplitter
+        "custom": CustomSplitter,
+        "semantic": SemanticSentenceSplitter,     # 语义分割（按段落+句子）
+        "semantic_embedding": None,              # 需要特殊处理
+        "sentence": SentenceSplitter,              # 严格按单句分割
+        "paragraph": ParagraphSplitter,           # 按段落分割
     }
 
+    # 特殊处理 embedding 分割器
+    if method == "semantic_embedding":
+        # 提取 embedding 相关参数
+        embedding_provider = kwargs.pop('embedding_provider', None)
+        if embedding_provider is None:
+            # 如果没有提供 provider，使用默认配置
+            # 从 kwargs 中获取模型配置
+            provider = kwargs.pop('embedding_provider_type', 'openai')
+            api_key = kwargs.pop('embedding_api_key', '')
+            base_url = kwargs.pop('embedding_base_url', 'https://api.minimax.chat/v1')
+            model = kwargs.pop('embedding_model', 'text-embedding-3-small')
+
+            if api_key:
+                embedding_provider = create_embedding_provider(
+                    provider, api_key, base_url, model
+                )
+
+        # 创建分割器
+        if embedding_provider:
+            return SemanticEmbeddingSplitter(
+                embedding_provider=embedding_provider,
+                **kwargs
+            )
+        else:
+            # 没有 embedding provider，降级到 semantic
+            method = "semantic"
+
     splitter_class = splitters.get(method, RecursiveTextSplitter)
     return splitter_class(**kwargs)
+
+
+class SemanticSentenceSplitter(TextSplitter):
+    """语义分割器 - 按段落优先，其次按句子"""
+
+    def __init__(self, chunk_size: int = 500, overlap: int = 50):
+        super().__init__(chunk_size, overlap)
+        self.splitter = RecursiveCharacterTextSplitter(
+            chunk_size=chunk_size,
+            chunk_overlap=overlap,
+            separators=[
+                "\n\n",  # 段落分隔优先
+                "。",    # 中文句号
+                "！",    # 中文感叹号
+                "？",    # 中文问号
+                ". ",    # 英文句号
+                "! ",    # 英文感叹号
+                "? ",    # 英文问号
+                "\n",    # 换行
+                " ",     # 空格
+            ],
+            length_function=self._count_chars
+        )
+
+    def _count_chars(self, text: str) -> int:
+        chinese_chars = len(re.findall(r'[\u4e00-\u9fff]', text))
+        other_chars = len(re.sub(r'[\u4e00-\u9fff]', '', text))
+        return chinese_chars + int(other_chars * 1.5)
+
+    def split(self, text: str) -> List[Dict]:
+        chunks = self.splitter.split_text(text)
+        result = []
+        for i, chunk in enumerate(chunks):
+            result.append({
+                "index": i,
+                "content": chunk.strip(),
+                "word_count": len(chunk.split()),
+                "char_count": len(chunk)
+            })
+        return result
+
+
+class SentenceSplitter(TextSplitter):
+    """严格按单句分割 - 每个chunk就是一句话"""
+
+    def __init__(self, chunk_size: int = 200, overlap: int = 0):
+        super().__init__(chunk_size, overlap)
+        # 只按句子结束符分割，不合并
+        self.splitter = RecursiveCharacterTextSplitter(
+            chunk_size=chunk_size,
+            chunk_overlap=overlap,
+            separators=[
+                "。",    # 中文句号
+                "！",    # 中文感叹号
+                "？",    # 中文问号
+                ". ",    # 英文句号
+                "! ",    # 英文感叹号
+                "? ",    # 英文问号
+                "\n",    # 换行
+                " ",     # 空格
+            ],
+            length_function=lambda x: len(x)
+        )
+
+    def split(self, text: str) -> List[Dict]:
+        chunks = self.splitter.split_text(text)
+        result = []
+        for i, chunk in enumerate(chunks):
+            chunk = chunk.strip()
+            if chunk:  # 跳过空chunk
+                result.append({
+                    "index": i,
+                    "content": chunk,
+                    "word_count": len(chunk.split()),
+                    "char_count": len(chunk)
+                })
+        return result
+
+
+class ParagraphSplitter(TextSplitter):
+    """按段落分割 - 以空行分隔"""
+
+    def __init__(self, chunk_size: int = 2000, overlap: int = 100):
+        overlap = min(overlap, chunk_size // 2)  # overlap 不能超过 chunk_size
+        super().__init__(chunk_size, overlap)
+
+    def split(self, text: str) -> List[Dict]:
+        # 按空行分割段落
+        paragraphs = re.split(r'\n\s*\n', text)
+        result = []
+        current_chunk = ""
+        chunk_index = 0
+
+        for para in paragraphs:
+            para = para.strip()
+            if not para:
+                continue
+
+            # 如果单个段落超过chunk_size，递归分割
+            if len(para) > self.chunk_size:
+                if current_chunk:
+                    result.append({
+                        "index": chunk_index,
+                        "content": current_chunk.strip(),
+                        "word_count": len(current_chunk.split()),
+                        "char_count": len(current_chunk)
+                    })
+                    chunk_index += 1
+                    current_chunk = ""
+
+                # 递归处理大段落
+                sub_splitter = RecursiveCharacterTextSplitter(
+                    chunk_size=self.chunk_size,
+                    chunk_overlap=self.overlap,
+                    separators=["\n", "。", "！", "？", ". ", "! ", "? "]
+                )
+                sub_chunks = sub_splitter.split_text(para)
+                for sub in sub_chunks:
+                    result.append({
+                        "index": chunk_index,
+                        "content": sub.strip(),
+                        "word_count": len(sub.split()),
+                        "char_count": len(sub)
+                    })
+                    chunk_index += 1
+            else:
+                if len(current_chunk) + len(para) > self.chunk_size:
+                    if current_chunk:
+                        result.append({
+                            "index": chunk_index,
+                            "content": current_chunk.strip(),
+                            "word_count": len(current_chunk.split()),
+                            "char_count": len(current_chunk)
+                        })
+                        chunk_index += 1
+                        current_chunk = ""
+
+                current_chunk += para + "\n\n"
+
+        # 添加最后一个chunk
+        if current_chunk.strip():
+            result.append({
+                "index": chunk_index,
+                "content": current_chunk.strip(),
+                "word_count": len(current_chunk.split()),
+                "char_count": len(current_chunk)
+            })
+
+        return result

backend/pyproject.toml

@@ -0,0 +1,125 @@
+[build-system]
+requires = ["setuptools>=61.0", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "yg-dataset"
+version = "1.0.0"
+description = "Dataset Generation Platform API"
+readme = "README.md"
+requires-python = ">=3.11"
+license = {text = "MIT"}
+authors = [
+    {name = "YG-Dataset Team", email = "team@yg-dataset.com"}
+]
+keywords = ["dataset", "machine-learning", "llm", "data-generation"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Framework :: FastAPI",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+]
+
+dependencies = [
+    "fastapi>=0.115.0",
+    "uvicorn[standard]>=0.30.0",
+    "python-multipart>=0.0.9",
+    "sqlalchemy>=2.0.0",
+    "alembic>=1.13.0",
+    "pydantic>=2.0.0",
+    "pydantic-settings>=2.0.0",
+    "pdfplumber>=0.10.4",
+    "python-docx>=1.1.0",
+    "openpyxl>=3.1.2",
+    "pandas>=2.2.0",
+    "ebooklib>=0.5",
+    "PyMuPDF>=1.24.0",
+    "langchain>=0.3.0",
+    "langchain-community>=0.2.0",
+    "langchain-openai>=0.1.0",
+    "tiktoken>=0.7.0",
+    "python-dotenv>=1.0.0",
+    "python-dateutil>=2.8.2",
+    "httpx>=0.27.0",
+    "aiofiles>=23.2.1",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=8.0.0",
+    "pytest-asyncio>=0.23.0",
+    "pytest-cov>=4.1.0",
+    "ruff>=0.3.0",
+    "mypy>=1.8.0",
+    "pre-commit>=3.6.0",
+    "black>=24.2.0",
+]
+
+[tool.uv]
+dev-dependencies = [
+    "pytest>=8.0.0",
+    "pytest-asyncio>=0.23.0",
+    "pytest-cov>=4.1.0",
+    "ruff>=0.3.0",
+    "mypy>=1.8.0",
+]
+
+[tool.ruff]
+line-length = 100
+target-version = "py311"
+
+[tool.ruff.lint]
+select = [
+    "E",   # pycodestyle errors
+    "W",   # pycodestyle warnings
+    "F",   # pyflakes
+    "I",   # isort
+    "B",   # flake8-bugbear
+    "C4",  # flake8-comprehensions
+    "UP",  # pyupgrade
+]
+ignore = [
+    "E501",  # line too long (handled by formatter)
+    "B008",  # do not perform function calls in argument defaults
+]
+
+[tool.mypy]
+python_version = "3.11"
+warn_return_any = true
+warn_unused_configs = true
+disallow_untyped_defs = false
+disallow_incomplete_defs = false
+check_untyped_defs = true
+no_implicit_optional = true
+warn_redundant_casts = true
+warn_unused_ignores = true
+warn_no_return = true
+
+[tool.pytest.ini_options]
+asyncio_mode = "auto"
+testpaths = ["tests"]
+python_files = ["test_*.py"]
+python_classes = ["Test*"]
+python_functions = ["test_*"]
+addopts = "-v --tb=short"
+
+[tool.coverage.run]
+source = ["app"]
+omit = [
+    "*/tests/*",
+    "*/venv/*",
+]
+
+[tool.coverage.report]
+exclude_lines = [
+    "pragma: no cover",
+    "def __repr__",
+    "raise AssertionError",
+    "raise NotImplementedError",
+    "if __name__ == .__main__.:",
+]
+
+[project.scripts]
+yg-dataset = "app.main:main"

backend/uploads/.gitkeep

@@ -0,0 +1 @@
+# This file ensures the uploads directory is tracked in git

bug修改.md

@@ -1,20 +0,0 @@
-# Bug 修改记录
-
-## 2026-03-17
-
-### 初始项目创建
-- 创建 YG-Dataset 重构项目
-- 搭建 FastAPI + Vue 3 基础架构
-
----
-
-## 修复记录格式
-
-### 日期
-**问题描述:**
-**原因:**
-**修复方案:**
-
----
-
-*持续更新中...*

daily-work/2026-03-17.md

@@ -0,0 +1,97 @@
+# 工作日志 - 2026-03-17
+
+## 项目信息
+- 项目: YG-Datasets
+- 路径: /data/code/YG-Datasets
+
+## 工作摘要
+今日共完成 16 项任务
+
+## 详细记录
+
+### 1. 🟣 Feature 后端核心架构模块
+- 时间: 17:28
+- 文件: backend/app/core/auth.py, backend/app/core/crud.py, backend/app/core/exceptions.py, backend/app/core/logging.py
+- 描述: 添加认证模块、CRUD 基础操作、异常处理、日志模块
+
+### 2. 🟣 Feature 后端 API Schemas 定义
+- 时间: 17:28
+- 文件: backend/app/schemas/chunk.py, backend/app/schemas/dataset.py, backend/app/schemas/eval.py
+- 描述: 添加数据结构的 Schema 定义
+
+### 3. 🟣 Feature 前端 TypeScript 类型定义和组件
+- 时间: 17:28
+- 文件: frontend/src/api/index.ts, frontend/src/components/, frontend/src/types/
+- 描述: 添加 TypeScript API 客户端和组件
+
+### 4. 🟣 Feature 前端页面功能和 UI 优化
+- 时间: 17:29
+- 文件: frontend/src/views/ModelSettingsView.vue, frontend/src/views/HomeView.vue
+- 描述: 添加模型配置页面，优化项目列表和删除功能
+
+### 5. ✅ Change 项目配置文件
+- 时间: 17:29
+- 文件: backend/pyproject.toml, frontend/tsconfig.json
+- 描述: 添加项目配置文件
+
+### 6. ✅ Change 一键启动脚本
+- 时间: 17:29
+- 文件: start.sh
+- 描述: 添加一键启动脚本
+
+### 7. 🟣 Feature 后端 API 端点实现
+- 时间: 17:29
+- 文件: backend/app/api/v1/projects/__init__.py, backend/app/api/v1/datasets/__init__.py
+- 描述: 更新 API 端点实现
+
+### 8. 🟣 Feature 后端核心模块和文件处理
+- 时间: 17:30
+- 文件: backend/app/core/config.py, backend/app/main.py, backend/app/models/models.py
+- 描述: 更新核心模块和文件处理器
+
+### 9. ✅ Change 前端依赖和路由配置
+- 时间: 17:30
+- 文件: frontend/package.json, frontend/src/router/index.js, frontend/vite.config.js
+- 描述: 更新依赖和路由配置
+
+### 10. 🔄 Refactor 前端 API 客户端重构
+- 时间: 17:30
+- 文件: frontend/src/api/index.js, frontend/src/api/index.ts
+- 描述: 用 TypeScript 版本替换 JavaScript API 客户端
+
+### 11. 🔴 Bugfix 修复返回按钮白色背景遮挡
+- 时间: 17:35
+- 文件: frontend/src/views/ModelSettingsView.vue
+- 描述: 修复模型配置页面返回按钮 hover 时白色背景遮挡问题
+
+### 12. 🔴 Bugfix 修复数据库初始化问题
+- 时间: 22:40
+- 文件: backend/app/core/database.py, backend/app/main.py
+- 描述: 修复数据库表未创建的问题，添加 models 导入确保 Base.metadata 包含所有模型
+
+### 13. 🔴 Bugfix 修复 API 响应序列化错误
+- 时间: 22:42
+- 文件: backend/app/api/v1/models/__init__.py, backend/app/schemas/model.py
+- 描述: 修复 SQLAlchemy ORM 对象无法序列化为 JSON 的问题，使用 model_validate() 转换
+
+### 14. 🟣 Feature 添加供应商默认 API Base URL
+- 时间: 22:45
+- 文件: frontend/src/views/ModelSettingsView.vue
+- 描述: 为 MiniMax、GLM、OpenAI Compatible 三个供应商添加默认 API Base URL，自动填充
+
+### 15. 🟣 Feature 实现模型连接测试功能
+- 时间: 22:50
+- 文件: backend/app/api/v1/models/__init__.py, frontend/src/views/ModelSettingsView.vue, frontend/src/api/index.ts
+- 描述: 后端添加测试连接 API，前端调用并显示连接状态（已联通/未联通/待测试）
+
+### 16. 🟣 Feature 创建 git-commit skill
+- 时间: 22:55
+- 文件: /root/.claude/skills/git-commit/SKILL.md
+- 描述: 创建 Git 分批提交技能，自动分析 git 状态，按功能分组文件，生成规范提交信息
+
+---
+
+## 其他工作
+
+- ✅ Change: 前端 UI 样式调整 - 添加 Ant Design Vue 组件库，调整 Select 组件暗色样式
+- 📝 Git: 推送所有代码更改到远程仓库，共 10 个 commit

daily-work/2026-03-18.md

@@ -0,0 +1,73 @@
+# 工作日志 - 2026-03-18
+
+## 项目信息
+- 项目: YG-Datasets
+- 路径: /data/code/YG-Datasets
+
+## 工作摘要
+今日共完成 11 项任务
+
+## 详细记录
+
+### 1. 🟣 Feature 完善日志系统，支持按日期分目录存储
+- 时间: 10:44
+- 文件: backend/app/core/logging.py, backend/app/main.py
+- 描述: 日志系统支持按日期分目录存储，便于日志管理和分析
+
+### 2. 🟣 Feature 完善前端功能，添加爬虫页面和项目分页
+- 时间: 10:45
+- 文件: frontend/src/views/HomeView.vue, frontend/src/views/CrawlerView.vue
+- 描述: 新增爬虫页面、composables 工具函数、项目列表分页功能
+
+### 3. 🟣 Feature 新增 composables 工具函数和爬虫页面
+- 时间: 10:45
+- 文件: frontend/src/composables/index.ts, frontend/src/composables/useFormatters.ts, frontend/src/composables/useProjects.ts
+- 描述: 添加前端工具函数 composables，实现项目、模型、格式化等复用逻辑
+
+### 4. 🔴 Bugfix 修复文件上传后异步处理失败问题
+- 时间: 16:08
+- 文件: backend/app/api/v1/files/__init__.py, backend/app/core/database.py
+- 描述: 修复 async_session_maker 引用错误，确保文件异步处理正常执行
+
+### 5. 🟣 Feature 添加语义嵌入文本分割功能
+- 时间: 16:08
+- 文件: backend/app/services/text_splitter/semantic_embedding.py, backend/app/services/text_splitter/splitter.py
+- 描述: 实现基于语义嵌入的文本分割算法，支持更智能的文本分块
+
+### 6. 🟣 Feature 更新 API 支持语义分割和 embedding 配置
+- 时间: 16:08
+- 文件: backend/app/api/v1/chunks/__init__.py, backend/app/schemas/
+- 描述: 后端 API 支持语义分割模式和 embedding 参数配置
+
+### 7. ✅ Change 优化文件管理上传流程和 UI 体验
+- 时间: 16:08
+- 文件: frontend/src/views/project/FileManage.vue
+- 描述: 优化文件上传流程，添加上传状态轮询、空状态处理、动画效果优化
+
+### 8. 🔄 Refactor 更新项目视图和文本分割页面
+- 时间: 16:08
+- 文件: frontend/src/views/ProjectView.vue, frontend/src/views/project/TextSplit.vue
+- 描述: 重构项目视图移除返回首页按钮，优化 TextSplit 页面样式和交互逻辑
+
+### 9. 🧹 Chore 删除废弃文件
+- 时间: 16:08
+- 文件: "bug修复.md"
+- 描述: 清理废弃文件
+
+---
+
+## 附加工作（会话中完成，尚未提交）
+
+### 10. ✅ Change 评估管理界面样式与文件管理保持一致
+- 时间: 17:32
+- 文件: frontend/src/views/project/EvalManage.vue
+- 描述: 评估管理界面采用与文件管理一致的样式：统计卡片带 glow 效果，空状态轨道动画、表格布局多选功能
+
+### 11. ✅ Change 问答管理界面样式与文件管理保持一致
+- 时间: 17:40
+- 文件: frontend/src/views/project/QuestionManage.vue
+- 描述: 问答管理界面采用与文件管理一致的样式：渐变标题、统计卡片，空状态动画、表格多选批量操作
+
+---
+
+*生成时间: 2026-03-18 17:45*

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/73518a92-e0d0-4f0a-903c-3beab46e6bdf.md

@@ -0,0 +1,266 @@
+文件编号：YG-CMMI-CM-PD07
+发布日期：2023-06-30
+现行版本：1.3
+
+商密【中】
+
+基线库管理规范
+
+修订历史记录
+| 日期          | 版本       | 说明       | 作者/修改人  | 审核   | 批准   |
+| ----------- | -------- | -------- | ------- | ---- | ---- |
+| 2012-11-14  | 1.0  新增  |          | 吴建春     | 李锋   | 卢晓民  |
+|             | 在规范描述中增加 | EAM 产品的内 |         |      |      |
+| 2013-7-23   | 1.1      |          | 陈来方     | 卢晓民  | 周立   |
+容和入基线的标准邮件
+调整“适用部门”、“相关文档”
+| 2014-09-23  | 1.2  章节、根据最新组织机构调整本 |     | 吴建春  | 卢晓民  | 李美平  |
+| ----------- | ------------------- | --- | ---- | ---- | ---- |
+文档引用的部门名称
+修订页眉中的商标引用，更新规
+| 2023-06-30  | 1.3  |     | 李锋  | 刘娟  | 向万红  |
+| ----------- | ---- | --- | --- | --- | ---- |
+范描述
+|     |     |     |     |     |     |
+| --- | --- | --- | --- | --- | --- |
+|     |     |     |     |     |     |
+
+远光软件股份有限公司  发布
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第1页 共7页
+目录
+1. 引言 ............................................................................................................................................................................ 2
+1.1 目的 ................................................................................................................................................................... 2
+1.2 适用范围 ........................................................................................................................................................... 2
+1.2.1 适用部门 ................................................................................................................................................... 2
+1.2.2 适用业务 ................................................................................................................................................... 2
+1.3 术语和缩略语 ................................................................................................................................................... 2
+2. 角色与职责 ................................................................................................................................................................ 2
+3. 规范描述 .................................................................................................................................................................... 3
+3.1 纳入基线管理的工作产品范围 ........................................................................................................................ 3
+3.2 通知入基线角色 ............................................................................................................................................... 3
+3.3 通知入基线时机 ............................................................................................................................................... 4
+3.4 文档存储地址 ................................................................................................................................................... 4
+3.5 权限管理 ........................................................................................................................................................... 5
+3.6 入基线文档规范 ............................................................................................................................................... 5
+3.6.1 模板引用 ................................................................................................................................................... 5
+3.6.2 格式要求 ................................................................................................................................................... 6
+3.7 文档入基线及基线变更的充分条件 ................................................................................................................ 6
+3.7.1 XX资料已入基线的邮件 .......................................................................................................................... 6
+4. 相关文档 .................................................................................................................................................................... 7
+5. 参考资料 .................................................................................................................................................................... 7
+商密【中】
+
+|     |     |     | 文件编号  | YG-CMMI-CM-PD07  |
+| --- | --- | --- | ----- | ---------------- |
+远光软件股份有限公司
+|     |     |     | 发布日期  | 2014-09-23  |
+| --- | --- | --- | ----- | ----------- |
+|     |     |     | 现行版本  | 1.2         |
+基线库管理规范
+|     |     |     | 页次  | 第2页 共7页  |
+| --- | --- | --- | --- | -------- |
+
+引言
+1.
+目的
+1.1
+为加强对研发线各阶段所输出工件入配置基线库活动的管理，明确规定入基线库的工件范围、及工件
+入基线库的时效性，特补充本规范。
+1.2  适用范围
+1.2.1 适用部门
+适用于公司产品研发部门、全资子公司，控股子公司参考执行。
+1.2.2 适用业务
+各产品研发线，在各阶段输出需入基线库工件的管理。
+1.3  术语和缩略语
+术语/缩略语  解释
+| CM(Configuration Management： |     |   是软件工程中的一项规程，包括相关工具和应用技术(过程和方 |     |     |
+| ---------------------------- | --- | ------------------------------- | --- | --- |
+| 配置管理)                        |     | 法)，公司用它来管理软件资产变更。               |     |     |
+| BL (Base Line：基线)            |     |   软件开发过程中的里程碑，它以一或多个软件配置项的交付为标 |     |     |
+志。基线由已经通过正式评审和批准的某规约或产品组成，它因
+
+此可以作为进一步开发的基础，并且只能通过正式的变更控制过
+程才能够改变。
+| 工件  |     |   软件研发生命周期各阶段的工作产品  |     |     |
+| --- | --- | -------------------- | --- | --- |
+
+2.  角色与职责
+| 序号  | 角色  | 职责  |     |     |
+| --- | --- | --- | --- | --- |
+商密【中】
+
+|     |     |     | 文件编号  | YG-CMMI-CM-PD07  |
+| --- | --- | --- | ----- | ---------------- |
+远光软件股份有限公司
+|     |     |     | 发布日期  | 2014-09-23  |
+| --- | --- | --- | ----- | ----------- |
+|     |     |     | 现行版本  | 1.2         |
+基线库管理规范
+|     |     |     | 页次  | 第3页 共7页  |
+| --- | --- | --- | --- | -------- |
+
+  负责将评审通过之后的文件入基线库
+配置管理员
+| 1   |     |     |     |     |
+| --- | --- | --- | --- | --- |
+负责将变更后的文件重新入基线库
+
+  负责确定项目哪些工作产品需入基线库管理
+项 目经理
+| 2   |     |   负责管理并要求配置管理员按要求将各阶段工作产品入基线库  |     |     |
+| --- | --- | ------------------------------- | --- | --- |
+  负责申请项目开发立项、发版材料入基线
+需求\设计\开发\测试
+  负责参与工件入基线前的评审；
+| 3   |     |     |     |     |
+| --- | --- | --- | --- | --- |
+责任人
+负责申请各阶段工作产品入基线库
+
+  负责审计各阶段工作产品是否按要求入基线库
+QA工程师
+| 4   |     |     |     |     |
+| --- | --- | --- | --- | --- |
+  负责审计各工作产品的完整性和符合性
+规范描述
+3.
+3.1  纳入基线管理的工作产品范围
+|           | 工件  | 阶段    |     |     |
+| --------- | --- | ----- | --- | --- |
+| 产品需求说明书   |     | 需求阶段  |     |     |
+| 需求规格说明书   |     | 需求阶段  |     |     |
+| 详细需求说明书   |     | 需求阶段  |     |     |
+| 功能设计文档    |     | 设计阶段  |     |     |
+| 详细设计文档    |     | 设计阶段  |     |     |
+| 测试用例      |     | 测试阶段  |     |     |
+| 项目开发立项材料  |     | 立项阶段  |     |     |
+| 项目结项材料    |     | 结项阶段  |     |     |
+注：
+1）开发阶段的代码管理依据《YG-CMMI-CM-PD04 配置管理规范》执行。
+2）发版阶段：集成测试结束之后，配置项清单检查之前。
+通知入基线角色
+3.2
+
+| 工件       | 入基线角色  |     |     |     |
+| -------- | ------ | --- | --- | --- |
+| 产品需求说明书  | 需求责任人  |     |     |     |
+商密【中】
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第4页 共7页
+需求规格说明书 需求责任人
+详细需求说明书 需求责任人
+功能设计文档 详细设计人员
+详细设计文档 详细设计人员
+测试用例 测试责任人
+项目开发立项材料 项目经理
+项目结项材料 项目经理
+3.3 通知入基线时机
+工件 入基线时机
+产品需求说明书 产品需求评审之后
+需求规格说明书 设计评审之后
+详细需求说明书评审之
+详细需求说明书
+后
+功能设计文档 评审之后
+详细设计文档 评审之后
+测试用例 测试用例评审之后
+项目开发立项材料 立项评审之后
+项目结项材料 结项评审之后
+3.4 文档存储地址
+（1）需求阶段
+需求规格说明书
+http:// 10.50.0.13/FMISdoc//baseline/ 需求规格说明书
+（2）设计阶段
+功能设计、详细设计文档：
+商密【中】
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第5页 共7页
+http:// 10.50.0.13/FMISdoc//baseline/设计文档/
+（3）测试阶段
+测试用例
+测试阶段工件列入VSS管理
+（4）开发阶段的代码管理
+统一列入ClearCase管理，依照《YG-CMMI-CM-PD04 配置管理规范》执行
+（5）项目开发立项材料
+项目开发立项材料
+http:// 10.50.0.13/FMISdoc//baseline/XXX(项目名称)/ 项目开发立项材料
+（6）项目结项材料
+项目结项材料
+http:// 10.50.0.13/FMISdoc//baseline/XXX(项目名称)/ 项目结项材料
+注：
+上述地址仅供参考，项目文档存储地址地址以实际立项计划中的要求为准。
+3.5 权限管理
+只允许本项目配置管理员有权限新增、删除、修改基线库内本项目的文档。
+其他项目干系人只能拥有本工作范围内工件的读取权限。
+3.6 入基线文档规范
+3.6.1 模板引用
+各阶段工件需按照公司CMMI研发体系中所发布的相应标准模板编写。
+商密【中】
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第6页 共7页
+3.6.2 格式要求
+各个工件除了按照标准模板执行外，还需要遵循如下通用要求：
+1） 文档显示的默认比例必须为100%；
+2） 正式文档必须记录审核人、批准人；
+3） 各个模板中，相应章节若无内容时必须写“无”来表明，而不允许直接删除本章节或留空。
+3.7 文档入基线及基线变更的充分条件
+文档入基线或申请变更基线文件时，必须输入入基线申请（变更申请），并保留有审批记录。
+入基线申请（变更申请）和审批记录，必须以正式的方式提交，如：依照RTC流程在相应单据中填写
+提交、或正式邮件提交。一切口头的或者利用其他交流工具提交的申请，都视为不符合。
+3.7.1 XX资料已入基线的邮件
+发件人：项目级配置管理师
+收件人：编写人（变更处理人）
+抄送人：配置经理、项目经理、资料使用人、QA
+邮件标题：关于XX资料已入基线的通知
+附件：无
+邮件内容模板如下：
+编写人（变更处理人）：
+您好！
+XX资料于MM-DD（周X）入基线，获取地址如下：XXX。
+配置管理工程师：XXX
+20XX年X月X日
+商密【中】
+
+|     |     |     | 文件编号  | YG-CMMI-CM-PD07  |
+| --- | --- | --- | ----- | ---------------- |
+远光软件股份有限公司
+|     |     |     | 发布日期  | 2014-09-23  |
+| --- | --- | --- | ----- | ----------- |
+|     |     |     | 现行版本  | 1.2         |
+基线库管理规范
+|     |     |     | 页次  | 第7页 共7页  |
+| --- | --- | --- | --- | -------- |
+
+4.  相关文档
+《YG-CMMI-CM-PC04 配置管理过程》
+《YG-CMMI-CM-PD04 配置管理规范》
+《YG-CMMI-PP-TEMP02 项目计划模板》
+《YG-CMMI-CM-TEMP13 配置管理计划模板》
+参考资料
+5.
+| 名称  | 来源  | 版本/日期  |     |     |
+| --- | --- | ------ | --- | --- |
+|     |     |        |     |     |
+
+商密【中】

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/73518a92-e0d0-4f0a-903c-3beab46e6bdf_YG-CMMI-CM-PD07 基线库管理规范.pdf.md

@@ -0,0 +1,266 @@
+文件编号：YG-CMMI-CM-PD07
+发布日期：2023-06-30
+现行版本：1.3
+
+商密【中】
+
+基线库管理规范
+
+修订历史记录
+| 日期          | 版本       | 说明       | 作者/修改人  | 审核   | 批准   |
+| ----------- | -------- | -------- | ------- | ---- | ---- |
+| 2012-11-14  | 1.0  新增  |          | 吴建春     | 李锋   | 卢晓民  |
+|             | 在规范描述中增加 | EAM 产品的内 |         |      |      |
+| 2013-7-23   | 1.1      |          | 陈来方     | 卢晓民  | 周立   |
+容和入基线的标准邮件
+调整“适用部门”、“相关文档”
+| 2014-09-23  | 1.2  章节、根据最新组织机构调整本 |     | 吴建春  | 卢晓民  | 李美平  |
+| ----------- | ------------------- | --- | ---- | ---- | ---- |
+文档引用的部门名称
+修订页眉中的商标引用，更新规
+| 2023-06-30  | 1.3  |     | 李锋  | 刘娟  | 向万红  |
+| ----------- | ---- | --- | --- | --- | ---- |
+范描述
+|     |     |     |     |     |     |
+| --- | --- | --- | --- | --- | --- |
+|     |     |     |     |     |     |
+
+远光软件股份有限公司  发布
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第1页 共7页
+目录
+1. 引言 ............................................................................................................................................................................ 2
+1.1 目的 ................................................................................................................................................................... 2
+1.2 适用范围 ........................................................................................................................................................... 2
+1.2.1 适用部门 ................................................................................................................................................... 2
+1.2.2 适用业务 ................................................................................................................................................... 2
+1.3 术语和缩略语 ................................................................................................................................................... 2
+2. 角色与职责 ................................................................................................................................................................ 2
+3. 规范描述 .................................................................................................................................................................... 3
+3.1 纳入基线管理的工作产品范围 ........................................................................................................................ 3
+3.2 通知入基线角色 ............................................................................................................................................... 3
+3.3 通知入基线时机 ............................................................................................................................................... 4
+3.4 文档存储地址 ................................................................................................................................................... 4
+3.5 权限管理 ........................................................................................................................................................... 5
+3.6 入基线文档规范 ............................................................................................................................................... 5
+3.6.1 模板引用 ................................................................................................................................................... 5
+3.6.2 格式要求 ................................................................................................................................................... 6
+3.7 文档入基线及基线变更的充分条件 ................................................................................................................ 6
+3.7.1 XX资料已入基线的邮件 .......................................................................................................................... 6
+4. 相关文档 .................................................................................................................................................................... 7
+5. 参考资料 .................................................................................................................................................................... 7
+商密【中】
+
+|     |     |     | 文件编号  | YG-CMMI-CM-PD07  |
+| --- | --- | --- | ----- | ---------------- |
+远光软件股份有限公司
+|     |     |     | 发布日期  | 2014-09-23  |
+| --- | --- | --- | ----- | ----------- |
+|     |     |     | 现行版本  | 1.2         |
+基线库管理规范
+|     |     |     | 页次  | 第2页 共7页  |
+| --- | --- | --- | --- | -------- |
+
+引言
+1.
+目的
+1.1
+为加强对研发线各阶段所输出工件入配置基线库活动的管理，明确规定入基线库的工件范围、及工件
+入基线库的时效性，特补充本规范。
+1.2  适用范围
+1.2.1 适用部门
+适用于公司产品研发部门、全资子公司，控股子公司参考执行。
+1.2.2 适用业务
+各产品研发线，在各阶段输出需入基线库工件的管理。
+1.3  术语和缩略语
+术语/缩略语  解释
+| CM(Configuration Management： |     |   是软件工程中的一项规程，包括相关工具和应用技术(过程和方 |     |     |
+| ---------------------------- | --- | ------------------------------- | --- | --- |
+| 配置管理)                        |     | 法)，公司用它来管理软件资产变更。               |     |     |
+| BL (Base Line：基线)            |     |   软件开发过程中的里程碑，它以一或多个软件配置项的交付为标 |     |     |
+志。基线由已经通过正式评审和批准的某规约或产品组成，它因
+
+此可以作为进一步开发的基础，并且只能通过正式的变更控制过
+程才能够改变。
+| 工件  |     |   软件研发生命周期各阶段的工作产品  |     |     |
+| --- | --- | -------------------- | --- | --- |
+
+2.  角色与职责
+| 序号  | 角色  | 职责  |     |     |
+| --- | --- | --- | --- | --- |
+商密【中】
+
+|     |     |     | 文件编号  | YG-CMMI-CM-PD07  |
+| --- | --- | --- | ----- | ---------------- |
+远光软件股份有限公司
+|     |     |     | 发布日期  | 2014-09-23  |
+| --- | --- | --- | ----- | ----------- |
+|     |     |     | 现行版本  | 1.2         |
+基线库管理规范
+|     |     |     | 页次  | 第3页 共7页  |
+| --- | --- | --- | --- | -------- |
+
+  负责将评审通过之后的文件入基线库
+配置管理员
+| 1   |     |     |     |     |
+| --- | --- | --- | --- | --- |
+负责将变更后的文件重新入基线库
+
+  负责确定项目哪些工作产品需入基线库管理
+项 目经理
+| 2   |     |   负责管理并要求配置管理员按要求将各阶段工作产品入基线库  |     |     |
+| --- | --- | ------------------------------- | --- | --- |
+  负责申请项目开发立项、发版材料入基线
+需求\设计\开发\测试
+  负责参与工件入基线前的评审；
+| 3   |     |     |     |     |
+| --- | --- | --- | --- | --- |
+责任人
+负责申请各阶段工作产品入基线库
+
+  负责审计各阶段工作产品是否按要求入基线库
+QA工程师
+| 4   |     |     |     |     |
+| --- | --- | --- | --- | --- |
+  负责审计各工作产品的完整性和符合性
+规范描述
+3.
+3.1  纳入基线管理的工作产品范围
+|           | 工件  | 阶段    |     |     |
+| --------- | --- | ----- | --- | --- |
+| 产品需求说明书   |     | 需求阶段  |     |     |
+| 需求规格说明书   |     | 需求阶段  |     |     |
+| 详细需求说明书   |     | 需求阶段  |     |     |
+| 功能设计文档    |     | 设计阶段  |     |     |
+| 详细设计文档    |     | 设计阶段  |     |     |
+| 测试用例      |     | 测试阶段  |     |     |
+| 项目开发立项材料  |     | 立项阶段  |     |     |
+| 项目结项材料    |     | 结项阶段  |     |     |
+注：
+1）开发阶段的代码管理依据《YG-CMMI-CM-PD04 配置管理规范》执行。
+2）发版阶段：集成测试结束之后，配置项清单检查之前。
+通知入基线角色
+3.2
+
+| 工件       | 入基线角色  |     |     |     |
+| -------- | ------ | --- | --- | --- |
+| 产品需求说明书  | 需求责任人  |     |     |     |
+商密【中】
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第4页 共7页
+需求规格说明书 需求责任人
+详细需求说明书 需求责任人
+功能设计文档 详细设计人员
+详细设计文档 详细设计人员
+测试用例 测试责任人
+项目开发立项材料 项目经理
+项目结项材料 项目经理
+3.3 通知入基线时机
+工件 入基线时机
+产品需求说明书 产品需求评审之后
+需求规格说明书 设计评审之后
+详细需求说明书评审之
+详细需求说明书
+后
+功能设计文档 评审之后
+详细设计文档 评审之后
+测试用例 测试用例评审之后
+项目开发立项材料 立项评审之后
+项目结项材料 结项评审之后
+3.4 文档存储地址
+（1）需求阶段
+需求规格说明书
+http:// 10.50.0.13/FMISdoc//baseline/ 需求规格说明书
+（2）设计阶段
+功能设计、详细设计文档：
+商密【中】
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第5页 共7页
+http:// 10.50.0.13/FMISdoc//baseline/设计文档/
+（3）测试阶段
+测试用例
+测试阶段工件列入VSS管理
+（4）开发阶段的代码管理
+统一列入ClearCase管理，依照《YG-CMMI-CM-PD04 配置管理规范》执行
+（5）项目开发立项材料
+项目开发立项材料
+http:// 10.50.0.13/FMISdoc//baseline/XXX(项目名称)/ 项目开发立项材料
+（6）项目结项材料
+项目结项材料
+http:// 10.50.0.13/FMISdoc//baseline/XXX(项目名称)/ 项目结项材料
+注：
+上述地址仅供参考，项目文档存储地址地址以实际立项计划中的要求为准。
+3.5 权限管理
+只允许本项目配置管理员有权限新增、删除、修改基线库内本项目的文档。
+其他项目干系人只能拥有本工作范围内工件的读取权限。
+3.6 入基线文档规范
+3.6.1 模板引用
+各阶段工件需按照公司CMMI研发体系中所发布的相应标准模板编写。
+商密【中】
+
+文件编号 YG-CMMI-CM-PD07
+远光软件股份有限公司
+发布日期 2014-09-23
+现行版本 1.2
+基线库管理规范
+页次 第6页 共7页
+3.6.2 格式要求
+各个工件除了按照标准模板执行外，还需要遵循如下通用要求：
+1） 文档显示的默认比例必须为100%；
+2） 正式文档必须记录审核人、批准人；
+3） 各个模板中，相应章节若无内容时必须写“无”来表明，而不允许直接删除本章节或留空。
+3.7 文档入基线及基线变更的充分条件
+文档入基线或申请变更基线文件时，必须输入入基线申请（变更申请），并保留有审批记录。
+入基线申请（变更申请）和审批记录，必须以正式的方式提交，如：依照RTC流程在相应单据中填写
+提交、或正式邮件提交。一切口头的或者利用其他交流工具提交的申请，都视为不符合。
+3.7.1 XX资料已入基线的邮件
+发件人：项目级配置管理师
+收件人：编写人（变更处理人）
+抄送人：配置经理、项目经理、资料使用人、QA
+邮件标题：关于XX资料已入基线的通知
+附件：无
+邮件内容模板如下：
+编写人（变更处理人）：
+您好！
+XX资料于MM-DD（周X）入基线，获取地址如下：XXX。
+配置管理工程师：XXX
+20XX年X月X日
+商密【中】
+
+|     |     |     | 文件编号  | YG-CMMI-CM-PD07  |
+| --- | --- | --- | ----- | ---------------- |
+远光软件股份有限公司
+|     |     |     | 发布日期  | 2014-09-23  |
+| --- | --- | --- | ----- | ----------- |
+|     |     |     | 现行版本  | 1.2         |
+基线库管理规范
+|     |     |     | 页次  | 第7页 共7页  |
+| --- | --- | --- | --- | -------- |
+
+4.  相关文档
+《YG-CMMI-CM-PC04 配置管理过程》
+《YG-CMMI-CM-PD04 配置管理规范》
+《YG-CMMI-PP-TEMP02 项目计划模板》
+《YG-CMMI-CM-TEMP13 配置管理计划模板》
+参考资料
+5.
+| 名称  | 来源  | 版本/日期  |     |     |
+| --- | --- | ------ | --- | --- |
+|     |     |        |     |     |
+
+商密【中】

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/ab0c2357-aae9-4939-b869-1dec750a815e.md

@@ -0,0 +1,394 @@
+文件编号：YG-CMMI-CM-PD04
+发布日期：2023-06-30
+现行版本：2.2
+
+                                                                      商密【中】
+
+代码提交规范
+
+修订历史记录
+
+日期
+2013-08-5
+
+版本
+1.0  新增代码提交规范                    张金金
+
+批准
+李剑/陈明有  周立
+
+作者/修改人
+
+说明
+
+审核
+
+2014-03-24
+
+1.1
+
+2017-08-30
+
+1.2
+
+2022-09-9
+
+2.0
+
+2023-02-20
+
+2.1
+
+2023-06-30
+
+2.2
+
+张羡
+
+杨莹
+
+陈金银
+
+陈斯华
+
+李美平
+
+增加代码提交时需写注释，便于构
+建系统识别变更集
+增加提交代码总则，增加附注 ”GRIS
+模块代码复审说明及操作”
+修订 1.2.1 适用部门 ；
+新增 2 角色与职责；
+新增 3  权限管理；
+修订 4.1 代码提交原则；
+修订 4.3.1 RTC 提交代码规范；
+新增 4.3.3 GIT 提交代码规范。
+修订适用范围、代码提交原则、增加
+代码库明细表。
+修订页眉中的商标引用，删除 RTC 代
+码提交规范章节
+
+姚国全
+陈金银          刘发/王优
+
+陈金银         黄德海
+
+贾士中
+
+向万红
+
+姚国全
+
+李锋
+
+刘娟
+
+远光软件股份有限公司  发布
+
+远光软件股份有限公司
+
+  代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-06-30
+
+2.2
+第 1 页 共 6 页
+
+目    录
+
+1.  引言 ................................................................................................................................................... 2
+1.1.  目的 ............................................................................................................................................................2
+1.2.  适用范围 ....................................................................................................................................................2
+1.2.1.  适用部门 ................................................................................................................................................2
+1.2.2.  适应业务 ................................................................................................................................................2
+2. 角色与职责 ............................................................................................................................................... 2
+3.  权限管理 ........................................................................................................................................... 2
+3.1.  用户管理及授权原则 ........................................................................................................................................2
+3.2.  代码权限开通或关闭流程 ........................................................................................................................3
+4.  规范描述 ........................................................................................................................................... 4
+4.1.  代码提交原则 ............................................................................................................................................4
+4.2.  代码库管理工具的特性规范 ....................................................................................................................4
+CC（ClearCase）提交代码规范 ..........................................................................................................4
+4.2.1.
+GIT 提交代码规范 .................................................................................................................................5
+4.2.2.
+5.  相关附件 ........................................................................................................................................... 5
+6.  相关文档 ........................................................................................................................................... 5
+7.  附代码库明细表（供参考） .............................................................................................................. 5
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 2 页 共 6 页
+
+1.  引言
+
+1.1. 目的
+
+为统一规范代码库用户管理及授权控制，规范代码提交，保证代码的安全性、完整性、可追溯性，
+
+特制定本规范，用以保障代码管理工具 GIT 和 CC（ClearCase）等提交代码活动的有序开展。
+
+1.2. 适用范围
+
+1.2.1.  适用部门
+
+适用于公司产品研发部门、全资子公司，控股子公司参考执行。
+
+1.2.2.  适应业务
+
+项目代码交付。
+
+2.  角色与职责
+
+序号
+
+角色
+
+职责
+
+1
+
+2
+
+ 部门经理
+
+项目经理
+/开发经理
+
+  负责审批本部门员工（含借调）提起的跨项目组代码权限申请。
+  负责审批项目组内员工（含借调）提起的代码权限申请。
+  及时关闭项目组调出、离职员工的代码权限。
+
+3
+
+ 开发工程师
+
+4
+
+ 配置管理工程师
+
+  负责发起代码权限开通或关闭申请。
+  明确代码库权限开通或关闭范围。
+  了解代码库使用规范及要求。
+  借调员工加入项目组的接受项目组统一管理。
+  员工调岗、离职时需及时申请权限关闭。
+  负责代码库权限的设置。
+  协助项目成员正常使用代码库。
+  填写单据相关信息，更改单据状态。
+
+3.  权限管理
+
+3.1. 用户管理及授权原则
+1)  用户代码权限，须经过项目经理（开发经理）或者部门经理审批，在审批通过后方可设置权限。
+
+2)  开通代码权限的用户，须妥善保管好自己的帐号和密码，不得转借他人使用，避免由此带来的泄密
+
+及代码追溯困难等问题的风险。
+
+3)  开通了代码权限的用户，发现帐号/密码泄漏，须及时修改密码，设置的密码应符合公司安全规范。
+
+因不遵守规范造成严重后果的，将按公司相关要求处理。
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 3 页 共 6 页
+
+4)  开通了代码权限的用户，须严格按照代码提交规范要求及对应工具的操作手册进行代码提交，确保
+
+代码的正确性、完整性及可追溯性。
+
+5)  开通了代码权限的用户，在调岗、离职时，需将代码交付完整后，按流程申请权限关闭。
+
+3.2. 代码权限开通或关闭流程
+
+ 入口准则
+
+1
+输入
+1
+
+  需要代码库权限开通或关闭
+
+  无
+
+过程步骤
+
+1.1    申请人在公司研发管理平台上新建《代码库权限申请单》，提交代码权限开通或关闭申请。
+申请人在《代码库权限申请单》填写相关内容，包括代码库地址、权限说明、执行人及其
+它需说明的内容，申请人申请所在项目组代码库权限的由项目经理（开发经理）审批，申
+
+1.2
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 4 页 共 6 页
+
+请部门其他项目组代码库权限的由项目归属的部门经理审批。
+
+1.3    申请人将《代码库权限申请单》提交给部门经理、或项目经理（开发经理）审批。
+
+1.4    部门经理、或项目经理（开发经理）审批通过，转单给配置管理工程师进行权限设置。
+
+1.5    配置管理工程师按单据设置相应代码权限，并完善单据信息，变更单据状态。
+
+输出
+1
+
+  《代码库权限申请单》
+
+出口准则
+
+1
+
+  《代码库权限申请单》完整填写，权限已设置，单据内容已完善，单据状态已更改为结束。
+
+裁剪
+裁剪内容  不可裁剪
+裁剪准则  无
+
+4.  规范描述
+
+4.1. 代码提交原则
+
+1)  遵守已发布文档《YG-CMMI-CM-PD04 配置管理规范》中关于代码管理的所有原则。
+
+2)  依据研发管理平台单据交付代码，没纳入版本的需求单/工作单，不允许提交代码至代码库。
+
+3)  交付代码时必须按单号正确填写注释，格式如下：
+
+注释必须含研发管理平台单据号，遵循以下三种要求：
+格式 1：单号    示例  843186
+格式 2：单号+空格+注释，示例  843186  注释内容
+格式 3：一个变更集可以对应多个单号，需都写入注释中，示例  843186  843187  注释内容
+补充说明：
+因解决编译报错修改的代码，关联导致编译报错时的代码变更集单据。
+解决合并冲突产生的变更集，关联产生冲突时的代码变更集单据。
+
+4)  代码在本地构建通过后才能交付至代码库。
+
+5)  所有提交发布流（分支）、受控流（分支）的代码需经复审后才可正式提交。各代码流（或分支）
+的具体管理要求请遵照发布文档《YG-CMMI-CM-PD04 配置管理规范（试行）.pdf》中涉及的关于代
+码管理的所有原则。
+
+6)  公司产品研发部门所有代码库的发布流（分支）、受控流（分支），原则上由配置管理工程师统一
+
+创建和管理。
+
+7)  未按以上原则进行代码提交和管理，造成代码问题的，依据公司相关规定进行处理。
+
+4.2. 代码库管理工具的特性规范
+
+4.2.1.  CC（ClearCase）提交代码规范
+
+1)  按单据号在研发管理平台上申请权限，由研发管理部部门经理审批后开通权限。
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 5 页 共 6 页
+
+2)  开发工程师在 CC 中提交代码，选中要检出的文件，填写修改该文件的研发管理平台中单据号，如
+
+果多个文件需要同时修改，则选中多个文件，一起检出，填写一个单号即可。
+
+3)  每次创建单号后，CC 都会记录，提示框显示出来的单号如果正确，无需重新创建单号。如果没有，
+
+创建即可。同一个单号修改多个文件，无需多次创建同一单号。
+
+4)  详细的 Clearcase 操作指南请参见附件《YG-CMMI-CM-GD01 ClearcaseLT 客户端操作指南.doc》。
+
+4.2.2.  GIT 提交代码规范
+
+1)  GIT 代码库因其分支灵活的特点，无法进行统一管控，项目组因各种原因须建立个人分支，须邮件
+向项目经理申请，项目经理审批通过后方可建立，对建立的个人分支，配置管理工程师不纳入整体
+的代码流管理，项目组需妥善管控，并定期清理，由此类分支管理混乱造成代码问题及引发严重后
+果的，依据公司相关规定进行处理。
+
+2)  详细的 GIT 操作手册请参见附件《YG-CMMI-CM-GD05 Git 使用手册-开发工程师》。
+
+5.  相关附件
+
+《YG-CMMI-CM-GD04 Git 安装指南-开发工程师》
+
+《YG-CMMI-CM-GD05 Git 使用手册-开发工程师》
+
+《YG-CMMI-CM-GD06 GAP 模块代码复审操作说明》
+
+《RTC 操作手册-开发工程师》
+
+6.  相关文档
+
+《YG-CMMI-CM-PD04 配置管理规范》
+
+《YG-CMMI-CM-PD03 配置项标识规范》
+
+《YG-CMMI-CM-GD01 ClearcaseLT 客户端操作指南》
+
+7. 附代码库明细表（供参考）
+
+代码库明细表.xls
+
+商密【中】
+

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/ab0c2357-aae9-4939-b869-1dec750a815e_YG-CMMI-CM-PD10 代码提交规范.pdf.md

@@ -0,0 +1,394 @@
+文件编号：YG-CMMI-CM-PD04
+发布日期：2023-06-30
+现行版本：2.2
+
+                                                                      商密【中】
+
+代码提交规范
+
+修订历史记录
+
+日期
+2013-08-5
+
+版本
+1.0  新增代码提交规范                    张金金
+
+批准
+李剑/陈明有  周立
+
+作者/修改人
+
+审核
+
+说明
+
+2014-03-24
+
+1.1
+
+2017-08-30
+
+1.2
+
+2022-09-9
+
+2.0
+
+2023-02-20
+
+2.1
+
+2023-06-30
+
+2.2
+
+张羡
+
+杨莹
+
+陈金银
+
+陈斯华
+
+李美平
+
+增加代码提交时需写注释，便于构
+建系统识别变更集
+增加提交代码总则，增加附注 ”GRIS
+模块代码复审说明及操作”
+修订 1.2.1 适用部门 ；
+新增 2 角色与职责；
+新增 3  权限管理；
+修订 4.1 代码提交原则；
+修订 4.3.1 RTC 提交代码规范；
+新增 4.3.3 GIT 提交代码规范。
+修订适用范围、代码提交原则、增加
+代码库明细表。
+修订页眉中的商标引用，删除 RTC 代
+码提交规范章节
+
+姚国全
+陈金银          刘发/王优
+
+陈金银         黄德海
+
+贾士中
+
+向万红
+
+姚国全
+
+李锋
+
+刘娟
+
+远光软件股份有限公司  发布
+
+远光软件股份有限公司
+
+  代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-06-30
+
+2.2
+第 1 页 共 6 页
+
+目    录
+
+1.  引言 ................................................................................................................................................... 2
+1.1.  目的 ............................................................................................................................................................2
+1.2.  适用范围 ....................................................................................................................................................2
+1.2.1.  适用部门 ................................................................................................................................................2
+1.2.2.  适应业务 ................................................................................................................................................2
+2. 角色与职责 ............................................................................................................................................... 2
+3.  权限管理 ........................................................................................................................................... 2
+3.1.  用户管理及授权原则 ........................................................................................................................................2
+3.2.  代码权限开通或关闭流程 ........................................................................................................................3
+4.  规范描述 ........................................................................................................................................... 4
+4.1.  代码提交原则 ............................................................................................................................................4
+4.2.  代码库管理工具的特性规范 ....................................................................................................................4
+CC（ClearCase）提交代码规范 ..........................................................................................................4
+4.2.1.
+GIT 提交代码规范 .................................................................................................................................5
+4.2.2.
+5.  相关附件 ........................................................................................................................................... 5
+6.  相关文档 ........................................................................................................................................... 5
+7.  附代码库明细表（供参考） .............................................................................................................. 5
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 2 页 共 6 页
+
+1.  引言
+
+1.1. 目的
+
+为统一规范代码库用户管理及授权控制，规范代码提交，保证代码的安全性、完整性、可追溯性，
+
+特制定本规范，用以保障代码管理工具 GIT 和 CC（ClearCase）等提交代码活动的有序开展。
+
+1.2. 适用范围
+
+1.2.1.  适用部门
+
+适用于公司产品研发部门、全资子公司，控股子公司参考执行。
+
+1.2.2.  适应业务
+
+项目代码交付。
+
+2.  角色与职责
+
+序号
+
+角色
+
+职责
+
+1
+
+2
+
+ 部门经理
+
+项目经理
+/开发经理
+
+  负责审批本部门员工（含借调）提起的跨项目组代码权限申请。
+  负责审批项目组内员工（含借调）提起的代码权限申请。
+  及时关闭项目组调出、离职员工的代码权限。
+
+3
+
+ 开发工程师
+
+4
+
+ 配置管理工程师
+
+  负责发起代码权限开通或关闭申请。
+  明确代码库权限开通或关闭范围。
+  了解代码库使用规范及要求。
+  借调员工加入项目组的接受项目组统一管理。
+  员工调岗、离职时需及时申请权限关闭。
+  负责代码库权限的设置。
+  协助项目成员正常使用代码库。
+  填写单据相关信息，更改单据状态。
+
+3.  权限管理
+
+3.1. 用户管理及授权原则
+1)  用户代码权限，须经过项目经理（开发经理）或者部门经理审批，在审批通过后方可设置权限。
+
+2)  开通代码权限的用户，须妥善保管好自己的帐号和密码，不得转借他人使用，避免由此带来的泄密
+
+及代码追溯困难等问题的风险。
+
+3)  开通了代码权限的用户，发现帐号/密码泄漏，须及时修改密码，设置的密码应符合公司安全规范。
+
+因不遵守规范造成严重后果的，将按公司相关要求处理。
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 3 页 共 6 页
+
+4)  开通了代码权限的用户，须严格按照代码提交规范要求及对应工具的操作手册进行代码提交，确保
+
+代码的正确性、完整性及可追溯性。
+
+5)  开通了代码权限的用户，在调岗、离职时，需将代码交付完整后，按流程申请权限关闭。
+
+3.2. 代码权限开通或关闭流程
+
+ 入口准则
+
+1
+输入
+1
+
+  需要代码库权限开通或关闭
+
+  无
+
+过程步骤
+
+1.1    申请人在公司研发管理平台上新建《代码库权限申请单》，提交代码权限开通或关闭申请。
+申请人在《代码库权限申请单》填写相关内容，包括代码库地址、权限说明、执行人及其
+它需说明的内容，申请人申请所在项目组代码库权限的由项目经理（开发经理）审批，申
+
+1.2
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 4 页 共 6 页
+
+请部门其他项目组代码库权限的由项目归属的部门经理审批。
+
+1.3    申请人将《代码库权限申请单》提交给部门经理、或项目经理（开发经理）审批。
+
+1.4    部门经理、或项目经理（开发经理）审批通过，转单给配置管理工程师进行权限设置。
+
+1.5    配置管理工程师按单据设置相应代码权限，并完善单据信息，变更单据状态。
+
+输出
+1
+
+  《代码库权限申请单》
+
+出口准则
+
+1
+
+  《代码库权限申请单》完整填写，权限已设置，单据内容已完善，单据状态已更改为结束。
+
+裁剪
+裁剪内容  不可裁剪
+裁剪准则  无
+
+4.  规范描述
+
+4.1. 代码提交原则
+
+1)  遵守已发布文档《YG-CMMI-CM-PD04 配置管理规范》中关于代码管理的所有原则。
+
+2)  依据研发管理平台单据交付代码，没纳入版本的需求单/工作单，不允许提交代码至代码库。
+
+3)  交付代码时必须按单号正确填写注释，格式如下：
+
+注释必须含研发管理平台单据号，遵循以下三种要求：
+格式 1：单号    示例  843186
+格式 2：单号+空格+注释，示例  843186  注释内容
+格式 3：一个变更集可以对应多个单号，需都写入注释中，示例  843186  843187  注释内容
+补充说明：
+因解决编译报错修改的代码，关联导致编译报错时的代码变更集单据。
+解决合并冲突产生的变更集，关联产生冲突时的代码变更集单据。
+
+4)  代码在本地构建通过后才能交付至代码库。
+
+5)  所有提交发布流（分支）、受控流（分支）的代码需经复审后才可正式提交。各代码流（或分支）
+的具体管理要求请遵照发布文档《YG-CMMI-CM-PD04 配置管理规范（试行）.pdf》中涉及的关于代
+码管理的所有原则。
+
+6)  公司产品研发部门所有代码库的发布流（分支）、受控流（分支），原则上由配置管理工程师统一
+
+创建和管理。
+
+7)  未按以上原则进行代码提交和管理，造成代码问题的，依据公司相关规定进行处理。
+
+4.2. 代码库管理工具的特性规范
+
+4.2.1.  CC（ClearCase）提交代码规范
+
+1)  按单据号在研发管理平台上申请权限，由研发管理部部门经理审批后开通权限。
+
+商密【中】
+
+远光软件股份有限公司
+
+代码提交规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD04
+2023-04-06
+
+2.1
+第 5 页 共 6 页
+
+2)  开发工程师在 CC 中提交代码，选中要检出的文件，填写修改该文件的研发管理平台中单据号，如
+
+果多个文件需要同时修改，则选中多个文件，一起检出，填写一个单号即可。
+
+3)  每次创建单号后，CC 都会记录，提示框显示出来的单号如果正确，无需重新创建单号。如果没有，
+
+创建即可。同一个单号修改多个文件，无需多次创建同一单号。
+
+4)  详细的 Clearcase 操作指南请参见附件《YG-CMMI-CM-GD01 ClearcaseLT 客户端操作指南.doc》。
+
+4.2.2.  GIT 提交代码规范
+
+1)  GIT 代码库因其分支灵活的特点，无法进行统一管控，项目组因各种原因须建立个人分支，须邮件
+向项目经理申请，项目经理审批通过后方可建立，对建立的个人分支，配置管理工程师不纳入整体
+的代码流管理，项目组需妥善管控，并定期清理，由此类分支管理混乱造成代码问题及引发严重后
+果的，依据公司相关规定进行处理。
+
+2)  详细的 GIT 操作手册请参见附件《YG-CMMI-CM-GD05 Git 使用手册-开发工程师》。
+
+5.  相关附件
+
+《YG-CMMI-CM-GD04 Git 安装指南-开发工程师》
+
+《YG-CMMI-CM-GD05 Git 使用手册-开发工程师》
+
+《YG-CMMI-CM-GD06 GAP 模块代码复审操作说明》
+
+《RTC 操作手册-开发工程师》
+
+6.  相关文档
+
+《YG-CMMI-CM-PD04 配置管理规范》
+
+《YG-CMMI-CM-PD03 配置项标识规范》
+
+《YG-CMMI-CM-GD01 ClearcaseLT 客户端操作指南》
+
+7. 附代码库明细表（供参考）
+
+代码库明细表.xls
+
+商密【中】
+

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/c09cbc25-6023-452c-9230-c6d359240a17.md

@@ -0,0 +1,279 @@
+文件编号：YG-CMMI-CM-PD03
+发布日期：2023-06-30
+现行版本：1.2
+
+                                                                      商密【中】
+
+配置项标识规范
+
+修订历史记录
+
+日期
+
+版本
+
+2008-06-19
+
+1.0
+
+说明
+新增文档，用于规范配置项标识
+工作，以便进行配置管理
+
+2010-07-01
+2023-06-30
+
+1.1  增加商密级别
+1.2  修订页眉中的商标引用
+
+作者/修改人
+
+审核
+
+杨莹
+
+吴建春
+李锋
+
+周立
+
+周立
+刘娟
+
+批准
+
+周立
+
+周立
+向万红
+
+远光软件股份有限公司  发布
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03PD03
+2023-06-30
+
+V1.2
+第 1 页 共 4 页
+
+目    录
+
+1. 引言 .......................................................................................................................................................... 2
+1.1.  目的 ....................................................................................................................................................................2
+1.1.1.  适用部门 ........................................................................................................................................................2
+1.1.2.  适用业务 ........................................................................................................................................................2
+1.2.  概述 ....................................................................................................................................................................2
+1.3.  术语和缩略语 ....................................................................................................................................................2
+2. 角色与职责 ............................................................................................................................................... 2
+3. 规范描述 ................................................................................................................................................... 2
+3.1.  配置项命名 ........................................................................................................................................................2
+3.1.1.  配置项/单元命名规范 ...................................................................................................................................2
+3.1.2.  配置单元中的配置项命名 ............................................................................................................................3
+3.2.  配置库对象标识 ................................................................................................................................................3
+3.2.1.  版本管理 ........................................................................................................................................................3
+3.2.2.  基线管理 ........................................................................................................................................................3
+3.2.3.  项目 ................................................................................................................................................................3
+3.2.4.  开发流 ............................................................................................................................................................3
+3.2.5.  活动 ................................................................................................................................................................3
+3.2.6.  视图 ................................................................................................................................................................3
+3.2.7.  VOB .................................................................................................................................................................3
+3.2.8.  componet 组件 ...............................................................................................................................................3
+3.2.9.  文件及文件夹 ................................................................................................................................................4
+4. 相关文档 ................................................................................................................................................... 4
+5. 参考资料 ................................................................................................................................................... 4
+
+商密【中】
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03
+2010-07-01
+
+V1.1
+第 2 页 共 4 页
+
+1.  引言
+
+1.1. 目的
+
+对纳入配置库的配置项统一标识，便于管理与查找。配置项的标识是为配置管理所使用的，不要和
+
+文档编号混淆。
+
+1.1.1. 适用部门
+
+配置管理部门。
+
+1.1.2. 适用业务
+
+配置项的标识活动。
+
+1.2. 概述
+
+描述对项目或产品工件如何命名、标记和编号。标识方案涵盖硬件、系统软件以及产品目录结构中
+
+列出的所有应用开发工件，例如，计划、模型、组件、测试软件、结果和数据、可执行文件，等等。
+
+解释
+
+  由配置管理视为一个单一整体而进行处理的工作产品（例如：在
+软件生存周期各阶段所产生的各种形式和各种版本的文档、程
+序、数据等）以及完成工作产品所需的软件工具和支持系统。
+  软件开发过程中的里程碑，它以一或多个软件配置项的交付为标
+志。基线由已经通过正式评审和批准的某规约或产品组成，它因
+此可以作为进一步开发的基础，并且只能通过正式的变更控制过
+程才能够改变.
+
+1.3. 术语和缩略语
+
+术语/缩略语
+
+SCI(Software Configuration Item：
+软件配置项)
+
+BL (BaseLine：基线)
+
+2.  角色与职责
+
+无
+
+3.  规范描述
+
+主要从配置项和配置库对象两个方面来定义如何进行标识，从而使配置库整洁有序。
+
+3.1. 配置项命名
+
+3.1.1. 配置项/单元命名规范
+
+配置项/单元命名规范为：项目简称_配置项/单元名称
+
+项目简称：为客户简称-项目名称缩写，应该是长度 8  位以内的英文字母与连接符的组合。
+
+商密【中】
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03
+2010-07-01
+
+V1.1
+第 3 页 共 4 页
+
+(项目简称由项目经理提供)
+
+3.1.2. 配置单元中的配置项命名
+
+对于测试用例等配置单元，其中的配置项的命名与编号规范，可以根据项目实际情况做相应约定，
+
+要求是体现配置项内容和所属功能模块等信息。
+
+文档类的配置项，以文档名称进行标识，标识时注意唯一性与可追溯性；
+
+源码类的配置项，可以直接使用程序名称作为配置项标识。
+
+3.2. 配置库对象标识
+
+3.2.1. 版本管理
+
+配置项每check out/in  一次，自身的版本号就会升一，可以直接看到。不允许随便删除版本，尤其是
+
+已经打上label或者基线的版本。
+
+3.2.2. 基线管理
+
+参见文档《基线发布控制规程》
+
+3.2.3. 项目
+
+项目名字由项目经理提供，一般为项目简称。
+
+3.2.4. 开发流
+
+开发流的标识统一使用小写，命名约定：[主项目]_[用途标识]，例如 ygerp_3.1_intergration、
+
+ygerp_3.1_report、ygerp_3.1_release 等。
+
+3.2.5. 活动
+
+针对的单据类型：R 加需求单号，Y 加优化单号，W 加工作单号，其他的向组织级配置管理工程师
+
+申报后再决定。
+
+3.2.6. 视图
+
+所有人员建立视图均采用以下模式：
+域用户名＋流名称，如 admin_ ygerp_3.1_report，此为 admin 用户在 ygerp_3.1_report 上建立的视图，
+
+不鼓励在同一流上建立多个视图。
+
+3.2.7. VOB
+
+按所存贮的数据类型命名，如 A.分析软件
+
+3.2.8. componet 组件
+
+管理组件统一采用 CBL.项目名称。如 CBL. ygerp_3.1。普通组件按功能进行命名。
+
+商密【中】
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03
+2010-07-01
+
+V1.1
+第 4 页 共 4 页
+
+3.2.9. 文件及文件夹
+
+统一采用小写字母进行命名。
+
+4.  相关文档
+
+无
+
+5.  参考资料
+
+名称
+
+无
+
+来源
+
+版本/日期
+
+商密【中】
+

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/c09cbc25-6023-452c-9230-c6d359240a17_YG-CMMI-CM-PD03 配置项标识规范.pdf.md

@@ -0,0 +1,279 @@
+文件编号：YG-CMMI-CM-PD03
+发布日期：2023-06-30
+现行版本：1.2
+
+                                                                      商密【中】
+
+配置项标识规范
+
+修订历史记录
+
+日期
+
+版本
+
+2008-06-19
+
+1.0
+
+说明
+新增文档，用于规范配置项标识
+工作，以便进行配置管理
+
+2010-07-01
+2023-06-30
+
+1.1  增加商密级别
+1.2  修订页眉中的商标引用
+
+作者/修改人
+
+审核
+
+杨莹
+
+吴建春
+李锋
+
+周立
+
+周立
+刘娟
+
+批准
+
+周立
+
+周立
+向万红
+
+远光软件股份有限公司  发布
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03PD03
+2023-06-30
+
+V1.2
+第 1 页 共 4 页
+
+目    录
+
+1. 引言 .......................................................................................................................................................... 2
+1.1.  目的 ....................................................................................................................................................................2
+1.1.1.  适用部门 ........................................................................................................................................................2
+1.1.2.  适用业务 ........................................................................................................................................................2
+1.2.  概述 ....................................................................................................................................................................2
+1.3.  术语和缩略语 ....................................................................................................................................................2
+2. 角色与职责 ............................................................................................................................................... 2
+3. 规范描述 ................................................................................................................................................... 2
+3.1.  配置项命名 ........................................................................................................................................................2
+3.1.1.  配置项/单元命名规范 ...................................................................................................................................2
+3.1.2.  配置单元中的配置项命名 ............................................................................................................................3
+3.2.  配置库对象标识 ................................................................................................................................................3
+3.2.1.  版本管理 ........................................................................................................................................................3
+3.2.2.  基线管理 ........................................................................................................................................................3
+3.2.3.  项目 ................................................................................................................................................................3
+3.2.4.  开发流 ............................................................................................................................................................3
+3.2.5.  活动 ................................................................................................................................................................3
+3.2.6.  视图 ................................................................................................................................................................3
+3.2.7.  VOB .................................................................................................................................................................3
+3.2.8.  componet 组件 ...............................................................................................................................................3
+3.2.9.  文件及文件夹 ................................................................................................................................................4
+4. 相关文档 ................................................................................................................................................... 4
+5. 参考资料 ................................................................................................................................................... 4
+
+商密【中】
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03
+2010-07-01
+
+V1.1
+第 2 页 共 4 页
+
+1.  引言
+
+1.1. 目的
+
+对纳入配置库的配置项统一标识，便于管理与查找。配置项的标识是为配置管理所使用的，不要和
+
+文档编号混淆。
+
+1.1.1. 适用部门
+
+配置管理部门。
+
+1.1.2. 适用业务
+
+配置项的标识活动。
+
+1.2. 概述
+
+描述对项目或产品工件如何命名、标记和编号。标识方案涵盖硬件、系统软件以及产品目录结构中
+
+列出的所有应用开发工件，例如，计划、模型、组件、测试软件、结果和数据、可执行文件，等等。
+
+解释
+
+  由配置管理视为一个单一整体而进行处理的工作产品（例如：在
+软件生存周期各阶段所产生的各种形式和各种版本的文档、程
+序、数据等）以及完成工作产品所需的软件工具和支持系统。
+  软件开发过程中的里程碑，它以一或多个软件配置项的交付为标
+志。基线由已经通过正式评审和批准的某规约或产品组成，它因
+此可以作为进一步开发的基础，并且只能通过正式的变更控制过
+程才能够改变.
+
+1.3. 术语和缩略语
+
+术语/缩略语
+
+SCI(Software Configuration Item：
+软件配置项)
+
+BL (BaseLine：基线)
+
+2.  角色与职责
+
+无
+
+3.  规范描述
+
+主要从配置项和配置库对象两个方面来定义如何进行标识，从而使配置库整洁有序。
+
+3.1. 配置项命名
+
+3.1.1. 配置项/单元命名规范
+
+配置项/单元命名规范为：项目简称_配置项/单元名称
+
+项目简称：为客户简称-项目名称缩写，应该是长度 8  位以内的英文字母与连接符的组合。
+
+商密【中】
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03
+2010-07-01
+
+V1.1
+第 3 页 共 4 页
+
+(项目简称由项目经理提供)
+
+3.1.2. 配置单元中的配置项命名
+
+对于测试用例等配置单元，其中的配置项的命名与编号规范，可以根据项目实际情况做相应约定，
+
+要求是体现配置项内容和所属功能模块等信息。
+
+文档类的配置项，以文档名称进行标识，标识时注意唯一性与可追溯性；
+
+源码类的配置项，可以直接使用程序名称作为配置项标识。
+
+3.2. 配置库对象标识
+
+3.2.1. 版本管理
+
+配置项每check out/in  一次，自身的版本号就会升一，可以直接看到。不允许随便删除版本，尤其是
+
+已经打上label或者基线的版本。
+
+3.2.2. 基线管理
+
+参见文档《基线发布控制规程》
+
+3.2.3. 项目
+
+项目名字由项目经理提供，一般为项目简称。
+
+3.2.4. 开发流
+
+开发流的标识统一使用小写，命名约定：[主项目]_[用途标识]，例如 ygerp_3.1_intergration、
+
+ygerp_3.1_report、ygerp_3.1_release 等。
+
+3.2.5. 活动
+
+针对的单据类型：R 加需求单号，Y 加优化单号，W 加工作单号，其他的向组织级配置管理工程师
+
+申报后再决定。
+
+3.2.6. 视图
+
+所有人员建立视图均采用以下模式：
+域用户名＋流名称，如 admin_ ygerp_3.1_report，此为 admin 用户在 ygerp_3.1_report 上建立的视图，
+
+不鼓励在同一流上建立多个视图。
+
+3.2.7. VOB
+
+按所存贮的数据类型命名，如 A.分析软件
+
+3.2.8. componet 组件
+
+管理组件统一采用 CBL.项目名称。如 CBL. ygerp_3.1。普通组件按功能进行命名。
+
+商密【中】
+
+远光软件股份有限公司
+
+配置项标识规范
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-PD03
+2010-07-01
+
+V1.1
+第 4 页 共 4 页
+
+3.2.9. 文件及文件夹
+
+统一采用小写字母进行命名。
+
+4.  相关文档
+
+无
+
+5.  参考资料
+
+名称
+
+无
+
+来源
+
+版本/日期
+
+商密【中】
+

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/c1c4326a-89c6-4fdf-b816-58c64f90c778.md

@@ -0,0 +1,434 @@
+文件编号：YG-CMMI-CM-GD04
+发布日期：2023-06-30
+现行版本：1.1
+
+商密【中】
+
+关于云效代码提交管理指南
+
+修订历史记录
+
+日期
+2022-08-03
+
+版本
+1.0  新增
+
+说明
+
+作者/修改人
+王优、刘发、胡玲、张金金
+
+审核
+黄德海
+
+批准
+姚国全
+
+2023-06-30
+
+1.1
+
+修订页眉中的商
+标引用
+
+李锋
+
+刘娟
+
+向万红
+
+远光软件股份有限公司  发布
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 1 页 共 8 页
+
+目    录
+
+第一章  总则 ........................................................................................................................................... 2
+第一条  目的 ........................................................................................................................................................2
+第二条  适用范围 ................................................................................................................................................2
+第三条  术语和缩略语 ........................................................................................................................................2
+第四条  职责分工 ................................................................................................................................................2
+第二章  云效构建流程 ............................................................................................................................. 3
+第五条  代码提交审批流程 ................................................................................................................................4
+第六条  平台代码提交流程 ................................................................................................................................5
+第七条  产品代码提交流程 ................................................................................................................................7
+第三章  其他 ........................................................................................................................................... 8
+第八条  依赖包白名单 ........................................................................................................................................8
+第九条  安全代码扫描 ........................................................................................................................................8
+第四章  附则 ........................................................................................................................................... 8
+第十条  本指南由 DAP 研发中心负责修订、解释 ............................................................................................8
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 2 页 共 8 页
+
+第一章  总则
+
+第一条 目的
+
+为规范财务中台云效系统代码提交流程，促进各部门高效协作，提高研发项目组提交代码的质量和
+
+效率，特制定本指南。
+
+第二条 适用范围
+
+适用于财务中台产品代码提交云效系统涉及的各研发部门。
+
+第三条 术语和缩略语
+
+术语/缩略语
+
+解释
+
+  云效是阿里巴巴云原生时代一站式 BizDevOps 平台，A 集团引
+
+入使用，进行端到端的全流程管理。
+
+  研发项目组自研程序包。
+
+  第三方开源程序包或付费购买的程序包。
+
+云效
+
+二方包
+
+三方包
+
+第四条 职责分工
+
+序号
+
+角色
+
+职责
+
+1
+
+ 研发项目组
+
+1.  项目经理负责云效代码提交申请；
+
+2.  负责平台/产品功能测试，保证代码质量；
+
+3.  负责输出静态资源到指定目录；
+
+4.  项目组对应的部门经理负责审批代码提交申请。
+
+2
+
+DAP 研发中心-测
+试 及 交 付 中 心
+（下文简称：测
+
+1.  负责准备测试环境；
+
+2.  负责产品功能的联调测试和自动化测试；
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 3 页 共 8 页
+
+试及交付中心）
+
+3.  平台冒烟通过后，负责更新测试环境，组织产品冒烟。
+
+信创平台部-测试
+部（下文简称：
+平台测试部）
+
+DAP 研发中心-研
+发管理部-配置组
+（下文简称：配
+置组）
+
+电力行业技术支
+持中心（下文简
+称：技术支持中
+心）
+
+1.  负责准备测试环境；
+
+2.  负责平台功能的联调测试和自动化测试，测试通过后邮件知会相关干系
+
+人。
+
+1.  负责代码本地编译验证并提取依赖包；
+
+2.  负责上传代码、依赖包、静态资源及镜像依赖文件至云效；
+
+3.  负责云效编译并输出镜像。
+
+1.  负责提供部署环境需要的镜像依赖文件；
+
+2.  负责组织在仿真环境和生产环境开展镜像部署；
+
+3.  负责审批云效代码的提交。
+
+ 电力行业一部
+
+1.  负责组织在仿真环境和生产环境开展测试。
+
+信创平台部-公共
+服务部-安全实验
+室（下文简称：
+安全实验室）
+
+1.  负责对研发项目组提交云效的代码进行安全扫描。
+
+3
+
+4
+
+5
+
+6
+
+7
+
+第二章  云效构建流程
+
+代码提交到云效系统有两种更新方式：
+
+1、  增量更新：按需发起申请，适用于不影响其他组功能使用的代码提交；
+
+2、  全量更新：按迭代或版本进行更新，由技术支持中心发起。更新前必须通过系统测试和安全测
+
+试。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 4 页 共 8 页
+
+第五条 代码提交审批流程
+
+1、  此流程适用增量更新时使用，全量更新由技术支持中心根据具体问题判断，直接发起。
+
+2、  发起申请：由研发项目组的项目经理申请提交代码到云效。
+
+3、  部门审核：由项目组对应的部门经理判断新提交代码是否影响其他项目组重新编译，只有不影
+响或能明确列出影响范围更新清单并通过测试验证的，才能提交技术支持中心负责人审批；平
+台代码不仅需要判断对平台其他组是否有影响，还需判断对其他产品是否有影响（是否需要其
+他产品调整代码或重新编译），如有影响审批不通过；如需更新，由技术支持中心根据具体问
+题判断是否发起全量更新。
+
+4、  技术支持中心审批：由技术支持中心负责人根据更新时间、更新内容等确定是否审批，审批通
+
+过后方可启动提交代码。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 5 页 共 8 页
+
+5、  启动云效代码提交流程：审批通过后，由配置组提取代码、依赖包和静态资源，启动提交代码
+
+的工作。
+
+第六条 平台代码提交流程
+
+1、  提交代码：由平台研发项目组对准备提交云效的代码、依赖包、静态资源（需同步输出到指定
+目录）等进行测试验证，此过程需要本地编译构建成功，通过平台测试部的冒烟测试，同时通
+过测试及交付中心的冒烟测试。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 6 页 共 8 页
+
+平台静态资源路径：
+\\10.50.9.101\server\modular whgz\ECP\ECPV8.0.0\finalZipFile
+
+2、  编译构建：由配置组根据平台提供的代码（当前为 boot 包），进行纯环境编译，确认获取依赖
+包（二方包、三方包）。如果选择增量更新，由发起者提供当前变更二方包列表、三方包列表
+和静态资源列表，如果选择全量更新，则无需提供。
+
+3、  获取镜像依赖文件：由技术支持中心按部署环境提供 dockerfile、配置文件和基础镜像等依赖文
+
+件给配置组。
+
+4、  云效提交：由配置组根据获取的依赖包，清除云效 maven 库相关二方包，提交更新的二方包、
+
+三方包、静态资源和 boot 代码并通过云效编译。
+
+5、  生成并推送镜像：由配置组云效编译并获取镜像依赖文件打包，输出镜像，并将生成的镜像推
+
+送云效镜像库。
+
+6、  仿真环境镜像部署：由技术支持中心组织在仿真环境进行镜像部署。
+
+7、  仿真环境测试：由电力行业一部组织在仿真环境开展功能测试验证。
+
+8、  生产环境镜像部署：由技术支持中心负责将镜像部署到生产环境，并在生产环境进行镜像部署。
+
+9、  生产环境测试：由电力行业一部组织在生产环境开展功能测试验证。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 7 页 共 8 页
+
+第七条 产品代码提交流程
+
+1、  提交代码：由产品研发项目组对准备提交云效的代码、依赖包、静态资源（需同步输出到指定
+目录）进行测试验证。此过程需要本地编译构建成功，通过测试及交付中心的产品联调测试和
+自动化测试。
+
+产品静态资源路径：
+\\10.50.9.101\server\modular whgz\dap9.0.0master_sp\fmp_resource\finalZipFile
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 8 页 共 8 页
+
+2、  编译构建：由配置组根据项目组提交的代码，进行纯环境编译，确认获取依赖包（二方包、三
+方包）。如果选择增量更新，由发起者提供当前变更二方包列表、三方包列表、静态资源列表，
+如果选择全量更新，则无需提供。
+
+3、  获取镜像所需依赖文件：由技术支持中心按部署环境提供 dockerfile、配置文件和基础镜像等依
+
+赖文件给配置组。
+
+4、  云效提交：由配置组根据获取的依赖包，清除云效 maven 库相关二方包，提交更新二方包、三
+
+方包、静态资源和代码并通过云效编译。
+
+5、  生成并推送镜像：由配置组云效编译并获取镜像依赖文件打包，输出镜像，并将生成的镜像推
+
+送云效镜像库。
+
+6、  仿真环境镜像部署：由技术支持中心组织在仿真环境进行镜像部署。
+
+7、  仿真环境测试：由电力行业一部组织在仿真环境开展功能测试验证。
+
+8、  生产环境镜像部署：由技术支持中心负责将镜像部署到生产环境，并在生产环境进行镜像部署。
+
+9、  生产环境测试：由电力行业一部组织在生产环境开展功能测试验证。
+
+第三章  其他
+
+第八条 依赖包白名单
+
+项目组提交云效的代码，编译所需依赖包，必须在配置组提供的依赖包白名单中，并通过 maven 编
+
+译。
+
+第九条 安全代码扫描
+
+按生产环境更新要求，研发项目组提交云效的代码需通过安全实验室相关工具的安全扫描。
+
+第四章  附则
+
+第十条 本指南由 DAP 研发中心负责修订、解释
+
+商密【中】
+

data/05ae84ba-9df7-4c42-89ab-51f817e48a06/ready/c1c4326a-89c6-4fdf-b816-58c64f90c778_YG-CMMI-CM-GD04 关于云效代码提交管理指南.pdf.md

@@ -0,0 +1,434 @@
+文件编号：YG-CMMI-CM-GD04
+发布日期：2023-06-30
+现行版本：1.1
+
+商密【中】
+
+关于云效代码提交管理指南
+
+修订历史记录
+
+日期
+2022-08-03
+
+版本
+1.0  新增
+
+说明
+
+作者/修改人
+王优、刘发、胡玲、张金金
+
+审核
+黄德海
+
+批准
+姚国全
+
+2023-06-30
+
+1.1
+
+修订页眉中的商
+标引用
+
+李锋
+
+刘娟
+
+向万红
+
+远光软件股份有限公司  发布
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 1 页 共 8 页
+
+目    录
+
+第一章  总则 ........................................................................................................................................... 2
+第一条  目的 ........................................................................................................................................................2
+第二条  适用范围 ................................................................................................................................................2
+第三条  术语和缩略语 ........................................................................................................................................2
+第四条  职责分工 ................................................................................................................................................2
+第二章  云效构建流程 ............................................................................................................................. 3
+第五条  代码提交审批流程 ................................................................................................................................4
+第六条  平台代码提交流程 ................................................................................................................................5
+第七条  产品代码提交流程 ................................................................................................................................7
+第三章  其他 ........................................................................................................................................... 8
+第八条  依赖包白名单 ........................................................................................................................................8
+第九条  安全代码扫描 ........................................................................................................................................8
+第四章  附则 ........................................................................................................................................... 8
+第十条  本指南由 DAP 研发中心负责修订、解释 ............................................................................................8
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 2 页 共 8 页
+
+第一章  总则
+
+第一条 目的
+
+为规范财务中台云效系统代码提交流程，促进各部门高效协作，提高研发项目组提交代码的质量和
+
+效率，特制定本指南。
+
+第二条 适用范围
+
+适用于财务中台产品代码提交云效系统涉及的各研发部门。
+
+第三条 术语和缩略语
+
+术语/缩略语
+
+解释
+
+  云效是阿里巴巴云原生时代一站式 BizDevOps 平台，A 集团引
+
+入使用，进行端到端的全流程管理。
+
+  研发项目组自研程序包。
+
+  第三方开源程序包或付费购买的程序包。
+
+云效
+
+二方包
+
+三方包
+
+第四条 职责分工
+
+序号
+
+角色
+
+职责
+
+1
+
+ 研发项目组
+
+1.  项目经理负责云效代码提交申请；
+
+2.  负责平台/产品功能测试，保证代码质量；
+
+3.  负责输出静态资源到指定目录；
+
+4.  项目组对应的部门经理负责审批代码提交申请。
+
+2
+
+DAP 研发中心-测
+试 及 交 付 中 心
+（下文简称：测
+
+1.  负责准备测试环境；
+
+2.  负责产品功能的联调测试和自动化测试；
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 3 页 共 8 页
+
+试及交付中心）
+
+3.  平台冒烟通过后，负责更新测试环境，组织产品冒烟。
+
+信创平台部-测试
+部（下文简称：
+平台测试部）
+
+DAP 研发中心-研
+发管理部-配置组
+（下文简称：配
+置组）
+
+电力行业技术支
+持中心（下文简
+称：技术支持中
+心）
+
+1.  负责准备测试环境；
+
+2.  负责平台功能的联调测试和自动化测试，测试通过后邮件知会相关干系
+
+人。
+
+1.  负责代码本地编译验证并提取依赖包；
+
+2.  负责上传代码、依赖包、静态资源及镜像依赖文件至云效；
+
+3.  负责云效编译并输出镜像。
+
+1.  负责提供部署环境需要的镜像依赖文件；
+
+2.  负责组织在仿真环境和生产环境开展镜像部署；
+
+3.  负责审批云效代码的提交。
+
+ 电力行业一部
+
+1.  负责组织在仿真环境和生产环境开展测试。
+
+信创平台部-公共
+服务部-安全实验
+室（下文简称：
+安全实验室）
+
+1.  负责对研发项目组提交云效的代码进行安全扫描。
+
+3
+
+4
+
+5
+
+6
+
+7
+
+第二章  云效构建流程
+
+代码提交到云效系统有两种更新方式：
+
+1、  增量更新：按需发起申请，适用于不影响其他组功能使用的代码提交；
+
+2、  全量更新：按迭代或版本进行更新，由技术支持中心发起。更新前必须通过系统测试和安全测
+
+试。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 4 页 共 8 页
+
+第五条 代码提交审批流程
+
+1、  此流程适用增量更新时使用，全量更新由技术支持中心根据具体问题判断，直接发起。
+
+2、  发起申请：由研发项目组的项目经理申请提交代码到云效。
+
+3、  部门审核：由项目组对应的部门经理判断新提交代码是否影响其他项目组重新编译，只有不影
+响或能明确列出影响范围更新清单并通过测试验证的，才能提交技术支持中心负责人审批；平
+台代码不仅需要判断对平台其他组是否有影响，还需判断对其他产品是否有影响（是否需要其
+他产品调整代码或重新编译），如有影响审批不通过；如需更新，由技术支持中心根据具体问
+题判断是否发起全量更新。
+
+4、  技术支持中心审批：由技术支持中心负责人根据更新时间、更新内容等确定是否审批，审批通
+
+过后方可启动提交代码。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 5 页 共 8 页
+
+5、  启动云效代码提交流程：审批通过后，由配置组提取代码、依赖包和静态资源，启动提交代码
+
+的工作。
+
+第六条 平台代码提交流程
+
+1、  提交代码：由平台研发项目组对准备提交云效的代码、依赖包、静态资源（需同步输出到指定
+目录）等进行测试验证，此过程需要本地编译构建成功，通过平台测试部的冒烟测试，同时通
+过测试及交付中心的冒烟测试。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 6 页 共 8 页
+
+平台静态资源路径：
+\\10.50.9.101\server\modular whgz\ECP\ECPV8.0.0\finalZipFile
+
+2、  编译构建：由配置组根据平台提供的代码（当前为 boot 包），进行纯环境编译，确认获取依赖
+包（二方包、三方包）。如果选择增量更新，由发起者提供当前变更二方包列表、三方包列表
+和静态资源列表，如果选择全量更新，则无需提供。
+
+3、  获取镜像依赖文件：由技术支持中心按部署环境提供 dockerfile、配置文件和基础镜像等依赖文
+
+件给配置组。
+
+4、  云效提交：由配置组根据获取的依赖包，清除云效 maven 库相关二方包，提交更新的二方包、
+
+三方包、静态资源和 boot 代码并通过云效编译。
+
+5、  生成并推送镜像：由配置组云效编译并获取镜像依赖文件打包，输出镜像，并将生成的镜像推
+
+送云效镜像库。
+
+6、  仿真环境镜像部署：由技术支持中心组织在仿真环境进行镜像部署。
+
+7、  仿真环境测试：由电力行业一部组织在仿真环境开展功能测试验证。
+
+8、  生产环境镜像部署：由技术支持中心负责将镜像部署到生产环境，并在生产环境进行镜像部署。
+
+9、  生产环境测试：由电力行业一部组织在生产环境开展功能测试验证。
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 7 页 共 8 页
+
+第七条 产品代码提交流程
+
+1、  提交代码：由产品研发项目组对准备提交云效的代码、依赖包、静态资源（需同步输出到指定
+目录）进行测试验证。此过程需要本地编译构建成功，通过测试及交付中心的产品联调测试和
+自动化测试。
+
+产品静态资源路径：
+\\10.50.9.101\server\modular whgz\dap9.0.0master_sp\fmp_resource\finalZipFile
+
+商密【中】
+
+远光软件股份有限公司
+
+关于云效代码提交管理指南
+
+文件编号
+发布日期
+
+现行版本
+页    次
+
+YG-CMMI-CM-GD04
+2023-06-30
+
+V1.1
+第 8 页 共 8 页
+
+2、  编译构建：由配置组根据项目组提交的代码，进行纯环境编译，确认获取依赖包（二方包、三
+方包）。如果选择增量更新，由发起者提供当前变更二方包列表、三方包列表、静态资源列表，
+如果选择全量更新，则无需提供。
+
+3、  获取镜像所需依赖文件：由技术支持中心按部署环境提供 dockerfile、配置文件和基础镜像等依
+
+赖文件给配置组。
+
+4、  云效提交：由配置组根据获取的依赖包，清除云效 maven 库相关二方包，提交更新二方包、三
+
+方包、静态资源和代码并通过云效编译。
+
+5、  生成并推送镜像：由配置组云效编译并获取镜像依赖文件打包，输出镜像，并将生成的镜像推
+
+送云效镜像库。
+
+6、  仿真环境镜像部署：由技术支持中心组织在仿真环境进行镜像部署。
+
+7、  仿真环境测试：由电力行业一部组织在仿真环境开展功能测试验证。
+
+8、  生产环境镜像部署：由技术支持中心负责将镜像部署到生产环境，并在生产环境进行镜像部署。
+
+9、  生产环境测试：由电力行业一部组织在生产环境开展功能测试验证。
+
+第三章  其他
+
+第八条 依赖包白名单
+
+项目组提交云效的代码，编译所需依赖包，必须在配置组提供的依赖包白名单中，并通过 maven 编
+
+译。
+
+第九条 安全代码扫描
+
+按生产环境更新要求，研发项目组提交云效的代码需通过安全实验室相关工具的安全扫描。
+
+第四章  附则
+
+第十条 本指南由 DAP 研发中心负责修订、解释
+
+商密【中】
+

frontend/index.html

@@ -8,6 +8,6 @@
   </head>
   <body>
     <div id="app"></div>
-    <script type="module" src="/src/main.js"></script>
+    <script type="module" src="/src/main.ts"></script>
   </body>
 </html>

frontend/package.json

@@ -8,17 +8,22 @@
     "preview": "vite preview"
   },
   "dependencies": {
+    "@element-plus/icons-vue": "^2.3.0",
+    "@vueuse/core": "^11.0.0",
+    "ant-design-vue": "^4.2.6",
+    "axios": "^1.7.0",
+    "element-plus": "^2.8.0",
+    "pinia": "^2.2.0",
     "vue": "^3.5.0",
     "vue-router": "^4.4.0",
-    "pinia": "^2.2.0",
-    "element-plus": "^2.8.0",
-    "@element-plus/icons-vue": "^2.3.0",
-    "axios": "^1.7.0",
-    "@vueuse/core": "^11.0.0",
     "xlsx": "^0.18.5"
   },
   "devDependencies": {
-    "@vitejs/plugin-vue": "^5.0.0",
-    "vite": "^6.0.0"
+    "@vitejs/plugin-vue": "^5.2.4",
+    "sass": "^1.77.0",
+    "sass-embedded": "^1.98.0",
+    "typescript": "^5.9.3",
+    "vite": "^6.0.0",
+    "vue-tsc": "^3.2.5"
   }
 }

frontend/src/App.vue

@@ -33,12 +33,12 @@ const locale = ref(zhCn)
 @import url('https://fonts.googleapis.com/css2?family=Outfit:wght@300;400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap');
 
 :root {
-  /* Core Colors - Deep Space */
-  --bg-primary: #030407;
-  --bg-secondary: #0a0a0f;
-  --bg-tertiary: #12121a;
-  --bg-elevated: #1a1a24;
-  --bg-hover: #22222e;
+  /* Core Colors - Deep Space with Gradient */
+  --bg-primary: #0a0a14;
+  --bg-secondary: #12121f;
+  --bg-tertiary: #1a1a2a;
+  --bg-elevated: #222233;
+  --bg-hover: #2a2a3d;
 
   /* Accent - Cyan Violet */
   --accent-primary: #00d4ff;
@@ -69,8 +69,8 @@ const locale = ref(zhCn)
   /* Effects */
   --glow-primary: 0 0 30px rgba(0, 212, 255, 0.25);
   --glow-secondary: 0 0 30px rgba(124, 58, 237, 0.25);
-  --glass-bg: rgba(18, 18, 26, 0.6);
-  --glass-border: rgba(255, 255, 255, 0.06);
+  --glass-bg: rgba(26, 26, 42, 0.65);
+  --glass-border: rgba(255, 255, 255, 0.08);
 
   /* Spacing */
   --radius-sm: 6px;
@@ -110,6 +110,12 @@ html, body, #app {
   min-height: 100vh;
   position: relative;
   overflow: hidden;
+  background:
+    radial-gradient(ellipse 100% 80% at 50% 120%, rgba(99, 102, 241, 0.15), transparent 50%),
+    radial-gradient(ellipse 80% 60% at 80% 10%, rgba(0, 212, 255, 0.12), transparent 40%),
+    radial-gradient(ellipse 70% 50% at 10% 80%, rgba(124, 58, 237, 0.1), transparent 40%),
+    radial-gradient(ellipse 60% 40% at 90% 90%, rgba(236, 72, 153, 0.08), transparent 40%),
+    linear-gradient(180deg, var(--bg-primary) 0%, #0f0f1a 100%);
 }
 
 .bg-mesh {
@@ -123,50 +129,66 @@ html, body, #app {
 .mesh-gradient {
   position: absolute;
   border-radius: 50%;
-  filter: blur(120px);
+  filter: blur(100px);
   opacity: 0.5;
-  animation: float 25s ease-in-out infinite;
+  animation: sciFiFloat 15s ease-in-out infinite;
 }
 
 .mesh-1 {
-  width: 700px;
-  height: 700px;
-  background: radial-gradient(circle, rgba(0, 212, 255, 0.35) 0%, transparent 70%);
-  top: -250px;
-  right: -150px;
+  width: 900px;
+  height: 900px;
+  background: radial-gradient(circle, rgba(0, 212, 255, 0.5) 0%, rgba(6, 182, 212, 0.25) 40%, transparent 70%);
+  top: -400px;
+  left: -300px;
   animation-delay: 0s;
+  animation-duration: 18s;
 }
 
 .mesh-2 {
-  width: 600px;
-  height: 600px;
-  background: radial-gradient(circle, rgba(124, 58, 237, 0.3) 0%, transparent 70%);
-  bottom: -200px;
-  left: -150px;
-  animation-delay: -8s;
+  width: 800px;
+  height: 800px;
+  background: radial-gradient(circle, rgba(124, 58, 237, 0.5) 0%, rgba(139, 92, 246, 0.25) 40%, transparent 70%);
+  bottom: -300px;
+  right: -200px;
+  animation-delay: -5s;
+  animation-duration: 20s;
 }
 
 .mesh-3 {
-  width: 500px;
-  height: 500px;
-  background: radial-gradient(circle, rgba(6, 182, 212, 0.2) 0%, transparent 70%);
-  top: 40%;
-  left: 30%;
-  animation-delay: -16s;
+  width: 700px;
+  height: 700px;
+  background: radial-gradient(circle, rgba(236, 72, 153, 0.4) 0%, rgba(168, 85, 247, 0.2) 40%, transparent 70%);
+  top: 50%;
+  left: 50%;
+  transform: translate(-50%, -50%);
+  animation-delay: -10s;
+  animation-duration: 22s;
 }
 
-@keyframes float {
-  0%, 100% { transform: translate(0, 0) scale(1); }
-  25% { transform: translate(40px, -40px) scale(1.08); }
-  50% { transform: translate(-30px, 30px) scale(0.95); }
-  75% { transform: translate(-40px, -25px) scale(1.03); }
+@keyframes sciFiFloat {
+  0%, 100% {
+    transform: translate(0, 0) scale(1) rotate(0deg);
+    opacity: 0.4;
+  }
+  25% {
+    transform: translate(200px, -150px) scale(1.2) rotate(10deg);
+    opacity: 0.65;
+  }
+  50% {
+    transform: translate(-150px, 100px) scale(0.85) rotate(-5deg);
+    opacity: 0.5;
+  }
+  75% {
+    transform: translate(100px, 150px) scale(1.15) rotate(8deg);
+    opacity: 0.6;
+  }
 }
 
 .noise-overlay {
   position: absolute;
   inset: 0;
-  background-image: url("data:image/svg+xml,%3Csvg viewBox='0 0 256 256' xmlns='http://www.w3.org/2000/svg'%3E%3Cfilter id='noise'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.9' numOctaves='4' stitchTiles='stitch'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' filter='url(%23noise)'/%3E%3C/svg%3E");
-  opacity: 0.025;
+  background-image: url("data:image/svg+xml,%3Csvg viewBox='0 0 256 256' xmlns='http://www.w3.org/2000/svg'%3E%3Cfilter id='noise'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.8' numOctaves='4' stitchTiles='stitch'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' filter='url(%23noise)'/%3E%3C/svg%3E");
+  opacity: 0.035;
 }
 
 .app-content {
@@ -248,6 +270,36 @@ html, body, #app {
   background: var(--bg-tertiary);
 }
 
+/* Ant Design Vue Select Dropdown */
+.ant-select-dropdown {
+  background: var(--bg-elevated) !important;
+  border: 1px solid var(--border-subtle) !important;
+  border-radius: 8px !important;
+  box-shadow: 0 6px 16px rgba(0, 0, 0, 0.3) !important;
+}
+
+.ant-select-item {
+  color: var(--text-primary) !important;
+  border-radius: 6px !important;
+}
+
+.ant-select-item-option-active:not(.ant-select-item-option-disabled) {
+  background: var(--bg-hover) !important;
+}
+
+.ant-select-item-option-selected:not(.ant-select-item-option-disabled) {
+  background: var(--accent-primary-muted) !important;
+  color: var(--accent-primary) !important;
+}
+
+.ant-select-selection-item {
+  color: var(--text-primary) !important;
+}
+
+.ant-select-selection-placeholder {
+  color: var(--text-secondary) !important;
+}
+
 .el-textarea__inner {
   background: var(--bg-tertiary) !important;
   border: 1px solid var(--border-subtle) !important;

frontend/src/api/index.js

@@ -1,81 +0,0 @@
-import axios from 'axios'
-
-const request = axios.create({
-  baseURL: '/api/v1',
-  timeout: 60000
-})
-
-// Request interceptor
-request.interceptors.request.use(
-  config => {
-    return config
-  },
-  error => {
-    return Promise.reject(error)
-  }
-)
-
-// Response interceptor
-request.interceptors.response.use(
-  response => {
-    return response.data
-  },
-  error => {
-    const message = error.response?.data?.message || error.message || '请求失败'
-    console.error('API Error:', message)
-    return Promise.reject(error)
-  }
-)
-
-export const projectApi = {
-  list: () => request.get('/projects/'),
-  get: (id) => request.get(`/projects/${id}`),
-  create: (data) => request.post('/projects/', data),
-  update: (id, data) => request.put(`/projects/${id}`, data),
-  delete: (id) => request.delete(`/projects/${id}`)
-}
-
-export const fileApi = {
-  upload: (projectId, formData) =>
-    request.post(`/projects/${projectId}/files/upload`, formData, {
-      headers: { 'Content-Type': 'multipart/form-data' }
-    }),
-  list: (projectId) => request.get(`/projects/${projectId}/files/`),
-  get: (projectId, fileId) => request.get(`/projects/${projectId}/files/${fileId}`),
-  delete: (projectId, fileId) => request.delete(`/projects/${projectId}/files/${fileId}`)
-}
-
-export const chunkApi = {
-  split: (projectId, data) => request.post(`/projects/${projectId}/chunks/split`, data),
-  list: (projectId, params) => request.get(`/projects/${projectId}/chunks/`, { params }),
-  get: (projectId, chunkId) => request.get(`/projects/${projectId}/chunks/${chunkId}`),
-  update: (projectId, chunkId, data) => request.put(`/projects/${projectId}/chunks/${chunkId}`, data),
-  delete: (projectId, chunkId) => request.delete(`/projects/${projectId}/chunks/${chunkId}`)
-}
-
-export const questionApi = {
-  generate: (projectId, data) => request.post(`/projects/${projectId}/generate-questions`, data),
-  list: (projectId, params) => request.get(`/projects/${projectId}/chunks/${params.chunkId}/questions`),
-  update: (projectId, questionId, data) => request.put(`/projects/${projectId}/questions/${questionId}`, data),
-  delete: (projectId, questionId) => request.delete(`/projects/${projectId}/questions/${questionId}`)
-}
-
-export const datasetApi = {
-  list: (projectId) => request.get(`/projects/${projectId}/datasets/`),
-  create: (projectId, data) => request.post(`/projects/${projectId}/datasets/`, data),
-  get: (projectId, datasetId) => request.get(`/projects/${projectId}/datasets/${datasetId}`),
-  delete: (projectId, datasetId) => request.delete(`/projects/${projectId}/datasets/${datasetId}`),
-  export: (projectId, datasetId, data) =>
-    request.post(`/projects/${projectId}/datasets/${datasetId}/export`, data, {
-      responseType: 'blob'
-    })
-}
-
-export const evalApi = {
-  list: (projectId) => request.get(`/projects/${projectId}/eval-datasets/`),
-  create: (projectId, data) => request.post(`/projects/${projectId}/eval-datasets/`, data),
-  run: (projectId, evalId) => request.post(`/projects/${projectId}/eval-datasets/${evalId}/evaluate`),
-  getResults: (projectId, taskId) => request.get(`/projects/${projectId}/eval-tasks/${taskId}`)
-}
-
-export default request

frontend/src/core/api/index.ts

@@ -0,0 +1,122 @@
+import axios from 'axios'
+import type { AxiosInstance } from 'axios'
+import type { Project, ProjectCreate, ProjectUpdate, Model, ModelCreate } from '@/shared/types'
+
+const apiBaseURL = import.meta.env.VITE_API_BASE_URL
+  || (import.meta.env.PROD
+    ? '/api/v1'
+    : `${window.location.protocol}//${window.location.hostname}:8000/api/v1`)
+
+const request: AxiosInstance = axios.create({
+  baseURL: apiBaseURL,
+  timeout: 60000
+})
+
+// Request interceptor
+request.interceptors.request.use(
+  config => {
+    return config
+  },
+  error => {
+    return Promise.reject(error)
+  }
+)
+
+// Response interceptor
+request.interceptors.response.use(
+  response => {
+    const data = response.data
+    // Handle new ApiResponse format
+    if (data.success !== undefined) {
+      if (data.success) {
+        // Check if this is a paginated response by checking for pagination field
+        if (data.pagination) {
+          // Return full response with pagination for paginated endpoints
+          return {
+            items: data.data,
+            total: data.pagination.total,
+            page: data.pagination.page,
+            page_size: data.pagination.page_size,
+            total_pages: data.pagination.total_pages
+          }
+        }
+        return data.data // Return the actual data
+      } else {
+        return Promise.reject(new Error(data.message || data.error || '请求失败'))
+      }
+    }
+    return data
+  },
+  error => {
+    const message = error.response?.data?.message || error.message || '请求失败'
+    console.error('API Error:', message)
+    return Promise.reject(error)
+  }
+)
+
+export const projectApi = {
+  list: (params?: { page?: number; page_size?: number }) =>
+    request.get<{ items: Project[]; pagination: { total: number } }>('/projects', { params }),
+  get: (id: string) => request.get<Project>(`/projects/${id}`),
+  create: (data: ProjectCreate) => request.post<{ id: string }>('/projects', data),
+  update: (id: string, data: ProjectUpdate) => request.put<Project>(`/projects/${id}`, data),
+  delete: (id: string) => request.delete(`/projects/${id}`)
+}
+
+export const fileApi = {
+  upload: (projectId: string, formData: FormData) =>
+    request.post(`/projects/${projectId}/files/upload`, formData, {
+      headers: { 'Content-Type': 'multipart/form-data' }
+    }),
+  list: (projectId: string) => request.get(`/projects/${projectId}/files`),
+  get: (projectId: string, fileId: string) => request.get(`/projects/${projectId}/files/${fileId}`),
+  delete: (projectId: string, fileId: string) => request.delete(`/projects/${projectId}/files/${fileId}`)
+}
+
+export const chunkApi = {
+  split: (projectId: string, data: any) =>
+    request.post(`/projects/${projectId}/chunks/split`, data, {
+      timeout: 300000
+    }),
+  list: (projectId: string, params?: any) => request.get(`/projects/${projectId}/chunks`, { params }),
+  get: (projectId: string, chunkId: string) => request.get(`/projects/${projectId}/chunks/${chunkId}`),
+  update: (projectId: string, chunkId: string, data: any) => request.put(`/projects/${projectId}/chunks/${chunkId}`, data),
+  delete: (projectId: string, chunkId: string) => request.delete(`/projects/${projectId}/chunks/${chunkId}`)
+}
+
+export const questionApi = {
+  generate: (projectId: string, data: any) => request.post(`/projects/${projectId}/questions/generate`, data),
+  list: (projectId: string, params?: any) => request.get(`/projects/${projectId}/questions`, { params }),
+  update: (projectId: string, questionId: string, data: any) => request.put(`/projects/${projectId}/questions/${questionId}`, data),
+  delete: (projectId: string, questionId: string) => request.delete(`/projects/${projectId}/questions/${questionId}`)
+}
+
+export const datasetApi = {
+  list: (projectId: string) => request.get(`/projects/${projectId}/datasets`),
+  create: (projectId: string, data: any) => request.post(`/projects/${projectId}/datasets`, data),
+  get: (projectId: string, datasetId: string) => request.get(`/projects/${projectId}/datasets/${datasetId}`),
+  delete: (projectId: string, datasetId: string) => request.delete(`/projects/${projectId}/datasets/${datasetId}`),
+  export: (projectId: string, datasetId: string, data: any) =>
+    request.post(`/projects/${projectId}/datasets/${datasetId}/export`, data, {
+      responseType: 'blob'
+    })
+}
+
+export const evalApi = {
+  list: (projectId: string) => request.get(`/projects/${projectId}/eval-datasets/`),
+  create: (projectId: string, data: any) => request.post(`/projects/${projectId}/eval-datasets/`, data),
+  run: (projectId: string, evalId: string) => request.post(`/projects/${projectId}/eval-datasets/${evalId}/evaluate`),
+  getResults: (projectId: string, taskId: string) => request.get(`/projects/${projectId}/eval-tasks/${taskId}`)
+}
+
+export const modelApi = {
+  list: () => request.get<Model[]>('/models/'),
+  get: (id: string) => request.get<Model>(`/models/${id}`),
+  create: (data: ModelCreate) => request.post<{ id: string }>('/models/', data),
+  update: (id: string, data: Partial<Model>) => request.put<Model>(`/models/${id}`, data),
+  delete: (id: string) => request.delete(`/models/${id}`),
+  setDefault: (id: string) => request.post(`/models/${id}/set-default`),
+  test: (id: string) => request.post<{ success: boolean; message: string }>(`/models/${id}/test`)
+}
+
+export default request

frontend/src/core/router/index.ts

@@ -0,0 +1,67 @@
+import { createRouter, createWebHistory, type RouteRecordRaw } from 'vue-router'
+
+const routes: RouteRecordRaw[] = [
+  {
+    path: '/',
+    name: 'Home',
+    component: () => import('@/pages/HomePage.vue')
+  },
+  {
+    path: '/project/:id',
+    name: 'Project',
+    component: () => import('@/pages/ProjectPage.vue'),
+    children: [
+      {
+        path: '',
+        redirect: to => `/project/${String(to.params.id)}/files`
+      },
+      {
+        path: 'files',
+        name: 'ProjectFiles',
+        component: () => import('@/pages/ProjectFilePage.vue')
+      },
+      {
+        path: 'split',
+        name: 'ProjectSplit',
+        component: () => import('@/pages/ProjectTextSplitPage.vue')
+      },
+      {
+        path: 'questions',
+        name: 'ProjectQuestions',
+        component: () => import('@/pages/ProjectQuestionPage.vue')
+      },
+      {
+        path: 'datasets',
+        name: 'ProjectDatasets',
+        component: () => import('@/pages/ProjectDatasetPage.vue')
+      },
+      {
+        path: 'eval',
+        name: 'ProjectEval',
+        component: () => import('@/pages/ProjectEvalPage.vue')
+      },
+      {
+        path: 'settings',
+        name: 'ProjectSettings',
+        component: () => import('@/pages/ProjectSettingsPage.vue')
+      }
+    ]
+  },
+  {
+    path: '/models',
+    name: 'ModelSettings',
+    component: () => import('@/pages/ModelSettingsPage.vue')
+  },
+  {
+    path: '/crawler',
+    name: 'Crawler',
+    component: () => import('@/pages/CrawlerPage.vue')
+  }
+]
+
+const router = createRouter({
+  history: createWebHistory(),
+  routes
+})
+
+export default router

frontend/src/main.ts

@@ -2,10 +2,12 @@ import { createApp } from 'vue'
 import { createPinia } from 'pinia'
 import ElementPlus from 'element-plus'
 import 'element-plus/dist/index.css'
+import Antd from 'ant-design-vue'
+import 'ant-design-vue/dist/reset.css'
 import * as ElementPlusIconsVue from '@element-plus/icons-vue'
 
 import App from './App.vue'
-import router from './router'
+import router from './core/router'
 
 const app = createApp(App)
 
@@ -17,5 +19,6 @@ for (const [key, component] of Object.entries(ElementPlusIconsVue)) {
 app.use(createPinia())
 app.use(router)
 app.use(ElementPlus)
+app.use(Antd)
 
 app.mount('#app')

frontend/src/page-logic/ModelSettingsPage.ts

@@ -0,0 +1,260 @@
+import { defineComponent } from 'vue'
+import { ref, reactive, computed, onMounted } from 'vue'
+import { useRouter } from 'vue-router'
+import { ElMessage } from 'element-plus'
+import type { ModelConfig, ProviderOption, ModelCreate, ModelType } from '@/shared/types'
+import { modelApi } from '@/core/api'
+import { watch } from 'vue'
+
+export default defineComponent({
+  name: 'ModelSettingsView',
+  setup() {
+    const router = useRouter()
+    
+    // 状态
+    const loading = ref(false)
+    const submitting = ref(false)
+    const deleting = ref(false)
+    const showAddDialog = ref(false)
+    const deleteDialogVisible = ref(false)
+    const modelToDelete = ref<ModelConfig | null>(null)
+    const models = ref<ModelConfig[]>([])
+    
+    // 表单
+    const modelForm = reactive<ModelCreate>({
+      provider: 'minimax',
+      model_type: 'chat',
+      model_name: '',
+      api_key: '',
+      api_base: 'https://api.minimax.chat/v1',
+      is_default: false
+    })
+    
+    // 供应商默认 API 地址
+    const providerDefaultUrls: Record<string, string> = {
+      minimax: 'https://api.minimax.chat/v1',
+      glm: 'https://open.bigmodel.cn/api/paas/v4',
+      openai: 'https://api.openai.com/v1',
+      ali: 'https://dashscope.aliyuncs.com/compatible-mode/v1'
+    }
+    
+    type ProviderOptionItem = ProviderOption & {
+      desc: string
+    }
+    
+    type ModelTypeOption = {
+      value: ModelType
+      label: string
+      abbr: string
+      desc: string
+    }
+    
+    // 提供商
+    const providers: ProviderOptionItem[] = [
+      { value: 'minimax', label: 'MiniMax', abbr: 'MM', desc: '适合国内接入，默认官方端点' },
+      { value: 'glm', label: 'GLM', abbr: 'GL', desc: '智谱接口，兼容常见模型配置' },
+      { value: 'openai', label: 'OpenAI Compatible', abbr: 'OP', desc: '适配 OpenAI 及兼容协议服务' },
+      { value: 'ali', label: '阿里云百炼', abbr: 'AL', desc: '默认走 DashScope 兼容模式端点' }
+    ]
+    
+    const modelTypes: ModelTypeOption[] = [
+      { value: 'chat', label: 'Chat', abbr: 'CH', desc: '标准对话生成模型' },
+      { value: 'vlm', label: 'VLM', abbr: 'VL', desc: '视觉语言模型，适合图文输入' },
+      { value: 'embedding', label: 'Embedding', abbr: 'EM', desc: '文本向量化与语义检索' },
+      { value: 'rerank', label: 'Rerank', abbr: 'RR', desc: '重排模型，用于检索结果排序' }
+    ]
+    
+    const normalizeModelType = (modelType?: string, modelName?: string): ModelType => {
+      if (modelType && modelTypes.some(type => type.value === modelType) && modelType !== 'chat') {
+        return modelType as ModelType
+      }
+    
+      const normalizedName = (modelName || '').trim().toLowerCase()
+    
+      if (['rerank', 'bce-reranker', 'gte-rerank'].some(keyword => normalizedName.includes(keyword))) {
+        return 'rerank'
+      }
+    
+      if (
+        [
+          'embedding',
+          'embed',
+          'text-embedding',
+          'bge-',
+          'bge_m3',
+          'gte-',
+          'm3e',
+          'e5-',
+          'jina-embeddings'
+        ].some(keyword => normalizedName.includes(keyword))
+      ) {
+        return 'embedding'
+      }
+    
+      if (['vl', 'vision', 'visual', 'multimodal', 'qwen-vl', 'gpt-4o'].some(keyword => normalizedName.includes(keyword))) {
+        return 'vlm'
+      }
+    
+      return 'chat'
+    }
+    
+    // 监听 provider 变化，自动设置默认 API 地址
+    
+    watch(() => modelForm.provider, (newProvider) => {
+      if (providerDefaultUrls[newProvider]) {
+        modelForm.api_base = providerDefaultUrls[newProvider]
+      }
+    })
+    
+    // 方法
+    const goHome = () => router.push('/')
+    
+    const getProviderAbbr = (provider: string) => {
+      const p = providers.find(p => p.value === provider)
+      return p?.abbr || '?'
+    }
+    
+    const getModelTypeLabel = (modelType?: string, modelName?: string) => {
+      const item = modelTypes.find(type => type.value === normalizeModelType(modelType, modelName))
+      return item?.label || 'Chat'
+    }
+    
+    const fetchModels = async () => {
+      loading.value = true
+      try {
+        const res = await modelApi.list()
+        // Handle different response formats
+        if (Array.isArray(res)) {
+          models.value = res.map(model => ({
+            ...model,
+            model_type: normalizeModelType(model.model_type, model.model_name)
+          }))
+        } else if (res?.data && Array.isArray(res.data)) {
+          models.value = res.data.map((model: ModelConfig) => ({
+            ...model,
+            model_type: normalizeModelType(model.model_type, model.model_name)
+          }))
+        } else {
+          models.value = []
+        }
+      } catch (error: any) {
+        console.error('获取模型列表失败:', error)
+        ElMessage.error(error?.message || '加载失败')
+      } finally {
+        loading.value = false
+      }
+    }
+    
+    const openAddDialog = () => {
+      modelForm.provider = 'minimax'
+      modelForm.model_type = 'chat'
+      modelForm.model_name = ''
+      modelForm.api_key = ''
+      modelForm.api_base = providerDefaultUrls['minimax']
+      modelForm.is_default = false
+      showAddDialog.value = true
+    }
+    
+    const addModel = async () => {
+      if (!modelForm.model_name || !modelForm.api_key) {
+        ElMessage.warning('请填写模型名称和 API Key')
+        return
+      }
+    
+      submitting.value = true
+      try {
+        // Convert is_default from boolean to string
+        const data = {
+          provider: modelForm.provider,
+          model_type: modelForm.model_type,
+          model_name: modelForm.model_name,
+          api_key: modelForm.api_key,
+          api_base: modelForm.api_base,
+          is_default: modelForm.is_default ? 'true' : 'false'
+        }
+        await modelApi.create(data)
+        ElMessage.success('添加成功')
+        showAddDialog.value = false
+        fetchModels()
+      } catch (error: any) {
+        console.error('添加模型失败:', error)
+        ElMessage.error(error?.message || '添加失败')
+      } finally {
+        submitting.value = false
+      }
+    }
+    
+    const confirmDelete = (model: ModelConfig) => {
+      modelToDelete.value = model
+      deleteDialogVisible.value = true
+    }
+    
+    const handleDelete = async () => {
+      if (!modelToDelete.value?.id) return
+      deleting.value = true
+      try {
+        await modelApi.delete(modelToDelete.value.id)
+        ElMessage.success('删除成功')
+        deleteDialogVisible.value = false
+        modelToDelete.value = null
+        fetchModels()
+      } catch (error: any) {
+        console.error('删除模型失败:', error)
+        ElMessage.error(error?.message || '删除失败')
+      } finally {
+        deleting.value = false
+      }
+    }
+    
+    const testConnection = async (model: ModelConfig) => {
+      ElMessage.info(`正在测试 ${model.model_name}...`)
+      try {
+        const res = await modelApi.test(model.id)
+        // Update model connection status from response
+        const modelItem = models.value.find(m => m.id === model.id)
+        if (modelItem && res?.model) {
+          modelItem.connection_status = res.model.connection_status
+          if (res.test_result?.success) {
+            ElMessage.success('连接成功！')
+          } else {
+            ElMessage.error(res.test_result?.message || '连接失败')
+          }
+        }
+      } catch (error: any) {
+        console.error('测试连接失败:', error)
+        const modelItem = models.value.find(m => m.id === model.id)
+        if (modelItem) {
+          modelItem.connection_status = 'disconnected'
+        }
+        ElMessage.error(error?.message || '连接失败')
+      }
+    }
+    
+    onMounted(() => fetchModels())
+
+    return {
+      router,
+      loading,
+      submitting,
+      deleting,
+      showAddDialog,
+      deleteDialogVisible,
+      modelToDelete,
+      models,
+      modelForm,
+      providerDefaultUrls,
+      providers,
+      modelTypes,
+      normalizeModelType,
+      goHome,
+      getProviderAbbr,
+      getModelTypeLabel,
+      fetchModels,
+      openAddDialog,
+      addModel,
+      confirmDelete,
+      handleDelete,
+      testConnection
+    }
+  }
+})

frontend/src/page-logic/ProjectFilePage.ts

@@ -0,0 +1,377 @@
+import { defineComponent } from 'vue'
+import { ref, computed, onMounted } from 'vue'
+import { useRoute } from 'vue-router'
+import { ElMessage } from 'element-plus'
+import { fileApi } from '@/core/api'
+import DeleteDialog from '@/shared/components/common/DeleteDialog.vue'
+
+export default defineComponent({
+  name: 'FileManage',
+  components: { DeleteDialog },
+  setup() {
+    const route = useRoute()
+    const projectId = computed(() => route.params.id)
+    
+    const loading = ref(false)
+    const files = ref([])
+    const filterStatus = ref('')
+    const isInitialLoad = ref(true)
+    
+    const filteredFiles = computed(() => {
+      if (!filterStatus.value) return files.value
+      return files.value.filter(f => f.status === filterStatus.value)
+    })
+    const uploadDialogVisible = ref(false)
+    const uploading = ref(false)
+    const uploadRef = ref(null)
+    const fileList = ref([])
+    const deleteDialogVisible = ref(false)
+    const pendingDeleteFile = ref(null)
+    const deletingFile = ref(false)
+    
+    // Multi-select
+    const selectedFiles = ref([])
+    
+    const isAllSelected = computed(() => filteredFiles.value.length > 0 && selectedFiles.value.length === filteredFiles.value.length)
+    const selectedCount = computed(() => selectedFiles.value.length)
+    
+    const toggleSelectAll = () => {
+      if (isAllSelected.value) {
+        selectedFiles.value = []
+      } else {
+        selectedFiles.value = filteredFiles.value.map(f => f.id)
+      }
+    }
+    
+    const toggleSelect = (fileId) => {
+      const index = selectedFiles.value.indexOf(fileId)
+      if (index === -1) {
+        selectedFiles.value.push(fileId)
+      } else {
+        selectedFiles.value.splice(index, 1)
+      }
+    }
+    
+    const isSelected = (fileId) => selectedFiles.value.includes(fileId)
+    
+    const clearSelection = () => {
+      selectedFiles.value = []
+    }
+    
+    const batchDeleteDialogVisible = ref(false)
+    const batchDeleting = ref(false)
+    const batchDeleteFiles = ref([])
+    
+    const batchDelete = async () => {
+      if (selectedFiles.value.length === 0) return
+      batchDeleteFiles.value = files.value.filter(f => selectedFiles.value.includes(f.id))
+      batchDeleteDialogVisible.value = true
+    }
+    
+    const executeBatchDelete = async () => {
+      if (selectedFiles.value.length === 0) return
+      batchDeleting.value = true
+      try {
+        for (const fileId of selectedFiles.value) {
+          await fileApi.delete(projectId.value, fileId)
+        }
+        ElMessage.success(`已删除 ${selectedFiles.value.length} 个文件`)
+        selectedFiles.value = []
+        batchDeleteDialogVisible.value = false
+        fetchFiles()
+      } catch (error) {
+        ElMessage.error('删除失败')
+      } finally {
+        batchDeleting.value = false
+      }
+    }
+    
+    // Preview
+    const previewVisible = ref(false)
+    const previewFile = ref(null)
+    const previewContent = ref('')
+    const previewLoading = ref(false)
+    const previewMode = ref('source') // 'source' | 'markdown'
+    const isPdfPreview = ref(false)
+    const pdfDataUrl = ref('')
+    const previewError = ref('')
+    
+    const completedFiles = computed(() => files.value.filter(f => f.status === 'completed').length)
+    const processingFiles = computed(() => files.value.filter(f => f.status === 'processing' || f.status === 'pending'))
+    const failedFiles = computed(() => files.value.filter(f => f.status === 'failed').length)
+    
+    const fetchFiles = async () => {
+      const wasInitial = isInitialLoad.value
+      loading.value = true
+      try {
+        const res = await fileApi.list(projectId.value)
+        files.value = res || []
+      } catch (error) {
+        files.value = []
+      } finally {
+        loading.value = false
+        if (wasInitial) {
+          isInitialLoad.value = false
+        }
+      }
+    }
+    
+    const handleUpload = () => {
+      fileList.value = []
+      uploadDialogVisible.value = true
+    }
+    
+    const handleChange = (file, files) => { fileList.value = files }
+    const handleRemove = (file, files) => { fileList.value = files }
+    
+    const triggerUpload = () => {
+      const input = uploadRef.value?.$el?.querySelector('input')
+      if (input) {
+        input.click()
+      }
+    }
+    
+    const submitUpload = async () => {
+      if (fileList.value.length === 0) {
+        ElMessage.warning('请先选择文件')
+        return
+      }
+    
+      // 保存上传前的文件数量
+      const prevFileCount = files.value.length
+    
+      // 先关闭对话框
+      uploadDialogVisible.value = false
+      ElMessage.success('已开始上传文件')
+    
+      // 设置上传状态，防止显示空状态
+      uploading.value = true
+    
+      // 在后台逐个上传（不等待上传完成）
+      const uploadPromises = fileList.value.map(async (item) => {
+        try {
+          const formData = new FormData()
+          formData.append('file', item.raw)
+          await fileApi.upload(projectId.value, formData)
+        } catch (error) {
+          console.error('上传失败:', error)
+        }
+      })
+    
+      // 立即刷新文件列表，显示新增的文件（状态为 processing）
+      await fetchFiles()
+    
+      // 如果之前没有文件，需要等待上传的Promise完成后再刷新一次
+      if (prevFileCount === 0) {
+        await Promise.all(uploadPromises)
+        await fetchFiles()
+      }
+    
+      // 持续轮询文件列表，直到没有 processing 状态的文件
+      const pollInterval = setInterval(async () => {
+        await fetchFiles()
+        // 检查是否还有处理中的文件
+        const hasProcessing = files.value.some(f => f.status === 'processing')
+        if (!hasProcessing) {
+          clearInterval(pollInterval)
+          uploading.value = false
+        }
+      }, 2000)
+    
+      // 最多轮询60秒
+      setTimeout(() => {
+        clearInterval(pollInterval)
+        uploading.value = false
+      }, 60000)
+    }
+    
+    const handleDelete = async (file) => {
+      try {
+        deletingFile.value = true
+        await fileApi.delete(projectId.value, file.id)
+        ElMessage.success('删除成功')
+        fetchFiles()
+      } catch (error) {
+        ElMessage.error('删除失败')
+      } finally {
+        deletingFile.value = false
+        deleteDialogVisible.value = false
+        pendingDeleteFile.value = null
+      }
+    }
+    
+    const openDeleteDialog = (file) => {
+      pendingDeleteFile.value = file
+      deleteDialogVisible.value = true
+    }
+    
+    const confirmDeleteFile = async () => {
+      if (!pendingDeleteFile.value) return
+      await handleDelete(pendingDeleteFile.value)
+    }
+    
+    const handlePreview = async (file) => {
+      previewFile.value = file
+      previewVisible.value = true
+      previewContent.value = ''
+      previewError.value = ''
+      previewLoading.value = true
+      previewMode.value = 'source'
+    
+      try {
+        await loadPreviewContent()
+      } finally {
+        previewLoading.value = false
+      }
+    }
+    
+    const loadPreviewContent = async () => {
+      if (!previewFile.value) return
+      previewLoading.value = true
+      previewContent.value = ''
+      previewError.value = ''
+      isPdfPreview.value = false
+      pdfDataUrl.value = ''
+    
+      try {
+        const endpoint = previewMode.value === 'source' ? 'raw' : 'content'
+        const response = await fetch(`/api/v1/projects/${projectId.value}/files/${previewFile.value.id}/${endpoint}`)
+    
+        if (response.ok) {
+          const text = await response.text()
+          if (text.startsWith('data:application/pdf;base64,')) {
+            isPdfPreview.value = true
+            pdfDataUrl.value = text
+            previewContent.value = ''
+          } else {
+            previewContent.value = text
+          }
+        } else if (response.status === 404) {
+          previewError.value = previewMode.value === 'source'
+            ? '源文件不存在或已被删除'
+            : 'Markdown 内容不存在，请等待处理完成'
+        } else if (response.status === 500) {
+          previewError.value = '服务器内部错误，请稍后重试'
+        } else {
+          previewError.value = `加载失败 (${response.status})`
+        }
+      } catch (error) {
+        previewError.value = '网络错误，请检查网络连接'
+      } finally {
+        previewLoading.value = false
+      }
+    }
+    
+    const switchPreviewMode = async (mode) => {
+      previewMode.value = mode
+      await loadPreviewContent()
+    }
+    
+    const getFileIcon = (type) => {
+      const map = { pdf: 'Document', docx: 'Document', xlsx: 'Grid', csv: 'Document', epub: 'Notebook', md: 'Document', txt: 'Document' }
+      return map[type] || 'Document'
+    }
+    
+    const getTypeColor = (type) => {
+      const map = {
+        pdf: '#ef4444',
+        docx: '#3b82f6',
+        xlsx: '#22c55e',
+        csv: '#22c55e',
+        epub: '#f59e0b',
+        md: '#8b5cf6',
+        txt: '#6b7280'
+      }
+      return map[type] || '#6b7280'
+    }
+    
+    const getFileExt = (filename) => {
+      if (!filename) return ''
+      const ext = filename.split('.').pop()?.toLowerCase()
+      return ext ? '.' + ext : ''
+    }
+    
+    const getStatusText = (status) => {
+      const map = {
+        processing: '处理中',
+        completed: '已完成',
+        failed: '失败',
+        pending: '待处理'
+      }
+      return map[status] || '未知'
+    }
+    
+    const formatSize = (bytes) => {
+      if (!bytes) return '0 B'
+      const k = 1024
+      const sizes = ['B', 'KB', 'MB', 'GB']
+      const i = Math.floor(Math.log(bytes) / Math.log(k))
+      return parseFloat((bytes / Math.pow(k, i)).toFixed(1)) + ' ' + sizes[i]
+    }
+    
+    const formatDate = (date) => {
+      if (!date) return ''
+      return new Date(date).toLocaleDateString('zh-CN', { month: 'short', day: 'numeric', hour: '2-digit', minute: '2-digit' })
+    }
+    
+    onMounted(() => fetchFiles())
+
+    return {
+      route,
+      projectId,
+      loading,
+      files,
+      filterStatus,
+      isInitialLoad,
+      filteredFiles,
+      uploadDialogVisible,
+      uploading,
+      uploadRef,
+      fileList,
+      deleteDialogVisible,
+      pendingDeleteFile,
+      deletingFile,
+      selectedFiles,
+      isAllSelected,
+      selectedCount,
+      toggleSelectAll,
+      toggleSelect,
+      isSelected,
+      clearSelection,
+      batchDeleteDialogVisible,
+      batchDeleting,
+      batchDeleteFiles,
+      batchDelete,
+      executeBatchDelete,
+      previewVisible,
+      previewFile,
+      previewContent,
+      previewLoading,
+      previewMode,
+      isPdfPreview,
+      pdfDataUrl,
+      previewError,
+      completedFiles,
+      processingFiles,
+      failedFiles,
+      fetchFiles,
+      handleUpload,
+      handleChange,
+      handleRemove,
+      triggerUpload,
+      submitUpload,
+      handleDelete,
+      openDeleteDialog,
+      confirmDeleteFile,
+      handlePreview,
+      loadPreviewContent,
+      switchPreviewMode,
+      getFileIcon,
+      getTypeColor,
+      getFileExt,
+      getStatusText,
+      formatSize,
+      formatDate
+    }
+  }
+})

frontend/src/page-logic/ProjectQuestionPage.ts

@@ -0,0 +1,252 @@
+import { defineComponent } from 'vue'
+import { ref, reactive, computed, onMounted } from 'vue'
+import { useRoute } from 'vue-router'
+import { ElMessage } from 'element-plus'
+import { chunkApi, questionApi, modelApi } from '@/core/api'
+
+export default defineComponent({
+  name: 'QuestionManage',
+  setup() {
+    const route = useRoute()
+    const projectId = computed(() => route.params.id)
+    
+    const loading = ref(false)
+    const isInitialLoad = ref(true)
+    const generating = ref(false)
+    const questions = ref([])
+    const chunks = ref([])
+    const availableModels = ref([])
+    const showGenerateDialog = ref(false)
+    const filterStatus = ref('')
+    const chunkMap = ref({})
+    
+    const DEFAULT_GENERATE_PROMPT = '你是一名高质量中文问答数据构建助手。请基于给定 chunk 内容生成准确、自然、可用于训练的数据集问答对。问题必须清晰具体，答案必须直接来自内容或基于内容做合理概括，不要编造原文没有的信息，不要输出与目录、导航、页眉页脚、噪声文字相关的问题。'
+    
+    const generateConfig = reactive({
+      model_id: '',
+      chunk_ids: [],
+      count: 3,
+      dirty_data_filter: true,
+      thinking_mode: true,
+      preset_prompt: DEFAULT_GENERATE_PROMPT
+    })
+    
+    // Multi-select
+    const selectedQuestions = ref([])
+    
+    const filteredQuestions = computed(() => {
+      if (!filterStatus.value) return questions.value
+      return questions.value.filter(q => q.source === filterStatus.value)
+    })
+    
+    const generatedCount = computed(() => questions.value.filter(q => q.source === 'generated').length)
+    const manualCount = computed(() => questions.value.filter(q => q.source === 'manual').length)
+    const failedCount = computed(() => questions.value.filter(q => q.status === 'failed').length)
+    const generateModels = computed(() => {
+      return availableModels.value.filter(model => {
+        const type = normalizeModelType(model.model_type, model.model_name)
+        return type === 'chat' || type === 'vlm'
+      })
+    })
+    
+    const isAllSelected = computed(() => filteredQuestions.value.length > 0 && selectedQuestions.value.length === filteredQuestions.value.length)
+    const selectedCount = computed(() => selectedQuestions.value.length)
+    
+    const toggleSelectAll = () => {
+      if (isAllSelected.value) {
+        selectedQuestions.value = []
+      } else {
+        selectedQuestions.value = filteredQuestions.value.map(q => q.id)
+      }
+    }
+    
+    const toggleSelect = (id) => {
+      const index = selectedQuestions.value.indexOf(id)
+      if (index === -1) {
+        selectedQuestions.value.push(id)
+      } else {
+        selectedQuestions.value.splice(index, 1)
+      }
+    }
+    
+    const isSelected = (id) => selectedQuestions.value.includes(id)
+    
+    const clearSelection = () => {
+      selectedQuestions.value = []
+    }
+    
+    const batchDelete = async () => {
+      if (selectedQuestions.value.length === 0) return
+      try {
+        for (const id of selectedQuestions.value) {
+          await questionApi.delete(projectId.value, id)
+        }
+        ElMessage.success(`已删除 ${selectedQuestions.value.length} 个问题`)
+        selectedQuestions.value = []
+        fetchQuestions()
+      } catch (error) {
+        ElMessage.error('删除失败')
+      }
+    }
+    
+    const normalizeModelType = (modelType, modelName = '') => {
+      if (modelType && modelType !== 'chat') {
+        return modelType
+      }
+      const normalizedName = String(modelName).trim().toLowerCase()
+      if (['rerank', 'bce-reranker', 'gte-rerank'].some(keyword => normalizedName.includes(keyword))) return 'rerank'
+      if (['embedding', 'embed', 'text-embedding', 'bge-', 'gte-', 'm3e', 'e5-', 'jina-embeddings'].some(keyword => normalizedName.includes(keyword))) return 'embedding'
+      if (['vl', 'vision', 'visual', 'multimodal', 'qwen-vl', 'gpt-4o'].some(keyword => normalizedName.includes(keyword))) return 'vlm'
+      return 'chat'
+    }
+    
+    const getProviderLabel = (provider) => {
+      const map = {
+        openai: 'OpenAI Compatible',
+        minimax: 'MiniMax',
+        glm: 'GLM',
+        ali: '阿里云百炼'
+      }
+      return map[provider] || provider
+    }
+    
+    const fetchAvailableModels = async () => {
+      try {
+        const res = await modelApi.list()
+        availableModels.value = Array.isArray(res) ? res : (res?.data || [])
+        if (!generateConfig.model_id && generateModels.value.length) {
+          const defaultModel = generateModels.value.find(model => model.is_default === 'true') || generateModels.value[0]
+          generateConfig.model_id = defaultModel?.id || ''
+        }
+      } catch (error) {
+        availableModels.value = []
+      }
+    }
+    
+    const fetchAllChunks = async () => {
+      const allChunks = []
+      let page = 1
+      let total = 0
+      do {
+        const res = await chunkApi.list(projectId.value, { page, page_size: 100 })
+        const items = res.items || res.data || []
+        total = res.total || res.pagination?.total || items.length
+        allChunks.push(...items)
+        page += 1
+      } while (allChunks.length < total)
+      return allChunks
+    }
+    
+    const fetchQuestions = async () => {
+      const wasInitial = isInitialLoad.value
+      loading.value = true
+      try {
+        const [chunkList, questionRes] = await Promise.all([
+          fetchAllChunks(),
+          questionApi.list(projectId.value, { page: 1, page_size: 500 })
+        ])
+        chunks.value = chunkList
+        chunkMap.value = Object.fromEntries(chunkList.map(chunk => [chunk.id, chunk]))
+        questions.value = questionRes.items || questionRes.data || []
+      } catch (error) {
+        questions.value = []
+      } finally {
+        loading.value = false
+        if (wasInitial) {
+          isInitialLoad.value = false
+        }
+      }
+    }
+    
+    const handleGenerate = async () => {
+      if (generateConfig.chunk_ids.length === 0) {
+        ElMessage.warning('请选择文本块')
+        return
+      }
+      if (!generateConfig.model_id) {
+        ElMessage.warning('请选择生成模型')
+        return
+      }
+      generating.value = true
+      try {
+        await questionApi.generate(projectId.value, generateConfig)
+        ElMessage.success('问题生成任务已启动')
+        showGenerateDialog.value = false
+        setTimeout(fetchQuestions, 2000)
+      } catch (error) {
+        ElMessage.error('生成失败')
+      } finally {
+        generating.value = false
+      }
+    }
+    
+    const handleDelete = async (question) => {
+      try {
+        await questionApi.delete(projectId.value, question.id)
+        ElMessage.success('删除成功')
+        fetchQuestions()
+      } catch (error) {
+        ElMessage.error('删除失败')
+      }
+    }
+    
+    const getTypeColor = (type) => {
+      const map = { 'fact': '#22c55e', 'summary': '#818cf8', 'reasoning': '#f59e0b' }
+      return map[type] || '#818cf8'
+    }
+    
+    const getTypeName = (type) => {
+      const map = { 'fact': '事实性', 'summary': '总结性', 'reasoning': '推理性' }
+      return map[type] || type
+    }
+    
+    const getSourceName = (source) => {
+      const map = { 'generated': 'AI生成', 'manual': '手动', 'failed': '失败' }
+      return map[source] || source
+    }
+    
+    onMounted(() => {
+      fetchAvailableModels()
+      fetchQuestions()
+    })
+
+    return {
+      route,
+      projectId,
+      loading,
+      isInitialLoad,
+      generating,
+      questions,
+      chunks,
+      availableModels,
+      showGenerateDialog,
+      filterStatus,
+      chunkMap,
+      DEFAULT_GENERATE_PROMPT,
+      generateConfig,
+      selectedQuestions,
+      filteredQuestions,
+      generatedCount,
+      manualCount,
+      failedCount,
+      generateModels,
+      isAllSelected,
+      selectedCount,
+      toggleSelectAll,
+      toggleSelect,
+      isSelected,
+      clearSelection,
+      batchDelete,
+      normalizeModelType,
+      getProviderLabel,
+      fetchAvailableModels,
+      fetchAllChunks,
+      fetchQuestions,
+      handleGenerate,
+      handleDelete,
+      getTypeColor,
+      getTypeName,
+      getSourceName
+    }
+  }
+})

frontend/src/page-logic/ProjectTextSplitPage.ts

@@ -0,0 +1,839 @@
+import { defineComponent } from 'vue'
+import { ref, reactive, computed, onMounted, onUnmounted, watch } from 'vue'
+import { useRoute, useRouter } from 'vue-router'
+import { ElMessage } from 'element-plus'
+import { fileApi, chunkApi, modelApi, questionApi } from '@/core/api'
+import DeleteDialog from '@/shared/components/common/DeleteDialog.vue'
+
+export default defineComponent({
+  name: 'TextSplit',
+  components: { DeleteDialog },
+  setup() {
+    const router = useRouter()
+    const route = useRoute()
+    const projectId = computed(() => route.params.id)
+    
+    const loading = ref(false)
+    const splitting = ref(false)
+    const files = ref([])
+    const filterStatus = ref('')
+    const fileChunks = ref({})
+    const isInitialLoad = ref(true)
+    const availableModels = ref([])
+    const filePollingTimer = ref(null)
+    const generateDialogVisible = ref(false)
+    const generatingQuestions = ref(false)
+    
+    const DEFAULT_GENERATE_PROMPT = '你是一名高质量中文问答数据构建助手。请基于给定 chunk 内容生成准确、自然、可用于训练的数据集问答对。问题必须清晰具体，答案必须直接来自内容或基于内容做合理概括，不要编造原文没有的信息，不要输出与目录、导航、页眉页脚、噪声文字相关的问题。'
+    
+    // Multi-select
+    const selectedFiles = ref([])
+    const splitDialogVisible = ref(false)
+    
+    // Chunk Preview Dialog
+    const chunkPreviewVisible = ref(false)
+    const previewFile = ref(null)
+    const previewChunks = ref([])
+    const previewLoading = ref(false)
+    const savingChunks = ref(false)
+    const deletingChunkId = ref('')
+    const deletingFileChunksId = ref('')
+    const deleteDialogVisible = ref(false)
+    const deleteDialogMode = ref('')
+    const deleteDialogTarget = ref(null)
+    const previewSearch = ref('')
+    const previewFilter = ref('all')
+    const previewJumpInput = ref('')
+    const selectedPreviewChunkId = ref('')
+    
+    const isAllSelected = computed(() => filteredFiles.value.length > 0 && selectedFiles.value.length === filteredFiles.value.length)
+    const selectedCount = computed(() => selectedFiles.value.length)
+    
+    const filteredFiles = computed(() => {
+      if (!filterStatus.value) return files.value
+      return files.value.filter(f => {
+        if (filterStatus.value === 'completed') {
+          return fileChunks.value[f.id]
+        }
+        if (filterStatus.value === 'processing') {
+          return f.status === 'processing'
+        }
+        return true
+      })
+    })
+    
+    const toggleSelectAll = () => {
+      if (isAllSelected.value) {
+        selectedFiles.value = []
+      } else {
+        selectedFiles.value = filteredFiles.value.map(f => f.id)
+      }
+    }
+    
+    const toggleSelect = (fileId) => {
+      const index = selectedFiles.value.indexOf(fileId)
+      if (index === -1) {
+        selectedFiles.value.push(fileId)
+      } else {
+        selectedFiles.value.splice(index, 1)
+      }
+    }
+    
+    const isSelected = (fileId) => selectedFiles.value.includes(fileId)
+    
+    const clearSelection = () => {
+      selectedFiles.value = []
+    }
+    
+    const splitConfig = reactive({
+      method: 'recursive',
+      chunk_size: 500,
+      overlap: 50,
+      separator: '\n\n',
+      embedding_model_id: '',
+      similarity_threshold: 0.3,
+      min_chunk_size: 100,
+    })
+    
+    const generateConfig = reactive({
+      model_id: '',
+      dirty_data_filter: true,
+      thinking_mode: true,
+      preset_prompt: DEFAULT_GENERATE_PROMPT,
+      count: 3,
+    })
+    
+    const methods = [
+      { value: 'recursive', label: '递归字符', tag: '基础' },
+      { value: 'semantic', label: '句段优先', tag: '规则' },
+      { value: 'semantic_embedding', label: '语义嵌入', tag: 'API' },
+      { value: 'markdown_structure', label: 'Markdown', tag: '结构' },
+    ]
+    
+    const completedFiles = computed(() => {
+      return Object.keys(fileChunks.value).length
+    })
+    
+    const processingCount = computed(() => {
+      return files.value.filter(f => f.status === 'processing').length
+    })
+    
+    const totalChunks = computed(() => {
+      return Object.values(fileChunks.value).reduce((sum, count) => sum + count, 0)
+    })
+    
+    const normalizeModelType = (modelType, modelName = '') => {
+      if (modelType && modelType !== 'chat') {
+        return modelType
+      }
+    
+      const normalizedName = String(modelName).trim().toLowerCase()
+    
+      if (['rerank', 'bce-reranker', 'gte-rerank'].some(keyword => normalizedName.includes(keyword))) {
+        return 'rerank'
+      }
+    
+      if ([
+        'embedding',
+        'embed',
+        'text-embedding',
+        'bge-',
+        'bge_m3',
+        'gte-',
+        'm3e',
+        'e5-',
+        'jina-embeddings'
+      ].some(keyword => normalizedName.includes(keyword))) {
+        return 'embedding'
+      }
+    
+      if (['vl', 'vision', 'visual', 'multimodal', 'qwen-vl', 'gpt-4o'].some(keyword => normalizedName.includes(keyword))) {
+        return 'vlm'
+      }
+    
+      return 'chat'
+    }
+    
+    const embeddingModels = computed(() => {
+      return availableModels.value.filter(model => normalizeModelType(model.model_type, model.model_name) === 'embedding')
+    })
+    
+    const selectedEmbeddingModel = computed(() => {
+      return embeddingModels.value.find(model => model.id === splitConfig.embedding_model_id) || null
+    })
+    
+    const generateModels = computed(() => {
+      return availableModels.value.filter(model => {
+        const type = normalizeModelType(model.model_type, model.model_name)
+        return type === 'chat' || type === 'vlm'
+      })
+    })
+    
+    const getProviderLabel = (provider) => {
+      const providerMap = {
+        openai: 'OpenAI Compatible',
+        minimax: 'MiniMax',
+        glm: 'GLM',
+        ali: '阿里云百炼'
+      }
+      return providerMap[provider] || provider
+    }
+    
+    const fetchAllChunks = async () => {
+      const allChunks = []
+      let page = 1
+      let total = 0
+    
+      do {
+        const res = await chunkApi.list(projectId.value, { page, page_size: 100 })
+        const items = res.items || res.data || []
+        total = res.total || res.pagination?.total || items.length
+        allChunks.push(...items)
+        page += 1
+      } while (allChunks.length < total)
+    
+      return allChunks
+    }
+    
+    const goToModelSettings = () => {
+      splitDialogVisible.value = false
+      generateDialogVisible.value = false
+      router.push('/models')
+    }
+    
+    const fetchAvailableModels = async () => {
+      try {
+        const res = await modelApi.list()
+        if (Array.isArray(res)) {
+          availableModels.value = res
+        } else if (res?.data && Array.isArray(res.data)) {
+          availableModels.value = res.data
+        } else {
+          availableModels.value = []
+        }
+      } catch (error) {
+        console.error(error)
+        availableModels.value = []
+      }
+    }
+    
+    watch(embeddingModels, (models) => {
+      if (!models.length) {
+        splitConfig.embedding_model_id = ''
+        return
+      }
+    
+      if (!models.some(model => model.id === splitConfig.embedding_model_id)) {
+        const defaultModel = models.find(model => model.is_default === 'true') || models[0]
+        splitConfig.embedding_model_id = defaultModel?.id || ''
+      }
+    }, { immediate: true })
+    
+    watch(generateModels, (models) => {
+      if (!models.length) {
+        generateConfig.model_id = ''
+        return
+      }
+    
+      if (!models.some(model => model.id === generateConfig.model_id)) {
+        const defaultModel = models.find(model => model.is_default === 'true') || models[0]
+        generateConfig.model_id = defaultModel?.id || ''
+      }
+    }, { immediate: true })
+    
+    const fetchFiles = async () => {
+      const wasInitial = isInitialLoad.value
+      loading.value = true
+      try {
+        const res = await fileApi.list(projectId.value)
+        files.value = res || []
+        // 获取每个文件的 chunk 数量
+        await fetchChunksCount()
+      } catch (error) {
+        console.error(error)
+      } finally {
+        loading.value = false
+        if (wasInitial) {
+          isInitialLoad.value = false
+        }
+      }
+    }
+    
+    const fetchChunksCount = async () => {
+      const counts = {}
+      for (const file of files.value) {
+        try {
+          const res = await chunkApi.list(projectId.value, { file_id: file.id })
+          const chunkList = res.items || res || []
+          if (chunkList.length > 0) {
+            counts[file.id] = chunkList.length
+          }
+        } catch (e) {
+          console.error(e)
+        }
+      }
+      fileChunks.value = counts
+    }
+    
+    const openSplitDialog = () => {
+      if (selectedFiles.value.length === 0) {
+        ElMessage.warning('请先选择要分割的文件')
+        return
+      }
+      if (!availableModels.value.length) {
+        fetchAvailableModels()
+      }
+      splitDialogVisible.value = true
+    }
+    
+    const handleBatchSplit = async () => {
+      if (selectedFiles.value.length === 0) {
+        ElMessage.warning('请先选择文件')
+        return
+      }
+      if (splitConfig.method === 'semantic_embedding' && !selectedEmbeddingModel.value) {
+        ElMessage.warning('请先选择已配置的 embedding 模型')
+        return
+      }
+      if (splitConfig.method === 'semantic_embedding' && !selectedEmbeddingModel.value?.api_key) {
+        ElMessage.warning('当前 embedding 模型缺少 API Key，请先到模型配置补全')
+        return
+      }
+      splitting.value = true
+      splitDialogVisible.value = false
+    
+      const successFiles = []
+      const failedFiles = []
+    
+      try {
+        for (const fileId of selectedFiles.value) {
+          const file = files.value.find(item => item.id === fileId)
+          const payload = {
+            file_id: fileId,
+            method: splitConfig.method,
+            chunk_size: splitConfig.chunk_size,
+            overlap: splitConfig.overlap,
+            separator: splitConfig.separator,
+            similarity_threshold: splitConfig.similarity_threshold,
+            min_chunk_size: splitConfig.min_chunk_size,
+          }
+    
+          if (splitConfig.method === 'semantic_embedding' && selectedEmbeddingModel.value) {
+            payload.embedding_provider = selectedEmbeddingModel.value.provider
+            payload.embedding_api_key = selectedEmbeddingModel.value.api_key
+            payload.embedding_base_url = selectedEmbeddingModel.value.api_base
+            payload.embedding_model = selectedEmbeddingModel.value.model_name
+          }
+    
+          try {
+            await chunkApi.split(projectId.value, payload)
+            successFiles.push(file?.filename || fileId)
+          } catch (error) {
+            console.error(error)
+            failedFiles.push({
+              name: file?.filename || fileId,
+              message: error?.message || '分割失败'
+            })
+          }
+        }
+    
+        if (successFiles.length && !failedFiles.length) {
+          ElMessage.success(`已为 ${successFiles.length} 个文件启动后台分割任务`)
+        } else if (successFiles.length && failedFiles.length) {
+          ElMessage.warning(`已启动 ${successFiles.length} 个，失败 ${failedFiles.length} 个：${failedFiles[0].name} - ${failedFiles[0].message}`)
+        } else if (failedFiles.length) {
+          ElMessage.error(`分割失败：${failedFiles[0].name} - ${failedFiles[0].message}`)
+          return
+        }
+    
+        // 清除选择
+        selectedFiles.value = []
+    
+        fetchFiles()
+      } finally {
+        splitting.value = false
+      }
+    }
+    
+    const openGenerateDialog = () => {
+      if (completedFiles.value === 0) {
+        ElMessage.warning('没有已分割的文件可生成')
+        return
+      }
+      if (!availableModels.value.length) {
+        fetchAvailableModels()
+      }
+      generateDialogVisible.value = true
+    }
+    
+    const handleBatchGenerate = async () => {
+      if (!generateConfig.model_id) {
+        ElMessage.warning('请选择用于生成问答的大语言模型')
+        return
+      }
+    
+      generatingQuestions.value = true
+      try {
+        const allChunks = await fetchAllChunks()
+        if (!allChunks.length) {
+          ElMessage.warning('当前项目还没有可用文本块')
+          return
+        }
+    
+        await questionApi.generate(projectId.value, {
+          chunk_ids: allChunks.map(chunk => chunk.id),
+          model_id: generateConfig.model_id,
+          dirty_data_filter: generateConfig.dirty_data_filter,
+          thinking_mode: generateConfig.thinking_mode,
+          preset_prompt: generateConfig.preset_prompt,
+          count: generateConfig.count
+        })
+    
+        generateDialogVisible.value = false
+        ElMessage.success(`已为 ${allChunks.length} 个文本块启动后台问答生成任务`)
+      } catch (error) {
+        console.error(error)
+        ElMessage.error(error?.message || '问答生成启动失败')
+      } finally {
+        generatingQuestions.value = false
+      }
+    }
+    
+    // Chunk Preview Methods
+    const openChunkPreview = async (file) => {
+      previewFile.value = file
+      previewSearch.value = ''
+      previewFilter.value = 'all'
+      previewJumpInput.value = ''
+      selectedPreviewChunkId.value = ''
+      chunkPreviewVisible.value = true
+      await fetchPreviewChunks(file.id)
+    }
+    
+    const previewChunksWithIndex = computed(() => {
+      return previewChunks.value.map((chunk, index) => ({
+        ...chunk,
+        displayIndex: index + 1
+      }))
+    })
+    
+    const isChunkModified = (chunk) => chunk.editingContent !== chunk.content
+    
+    const filteredPreviewChunks = computed(() => {
+      const keyword = previewSearch.value.trim().toLowerCase()
+    
+      return previewChunksWithIndex.value.filter(chunk => {
+        const matchesFilter = previewFilter.value !== 'modified' || isChunkModified(chunk)
+        if (!matchesFilter) return false
+        if (!keyword) return true
+    
+        return String(chunk.displayIndex).includes(keyword) || chunk.content.toLowerCase().includes(keyword) || chunk.editingContent.toLowerCase().includes(keyword)
+      })
+    })
+    
+    const activePreviewChunk = computed(() => {
+      return previewChunks.value.find(chunk => chunk.id === selectedPreviewChunkId.value) || null
+    })
+    
+    const activePreviewChunkIndex = computed(() => {
+      const index = previewChunks.value.findIndex(chunk => chunk.id === selectedPreviewChunkId.value)
+      return index === -1 ? 0 : index + 1
+    })
+    
+    const modifiedPreviewCount = computed(() => {
+      return previewChunks.value.filter(chunk => isChunkModified(chunk)).length
+    })
+    
+    const deleteDialogTitle = computed(() => {
+      if (deleteDialogMode.value === 'chunk') return '删除分片'
+      if (deleteDialogMode.value === 'file-chunks') return '删除全部分块'
+      return '删除'
+    })
+    
+    const deleteDialogItemName = computed(() => {
+      if (deleteDialogMode.value === 'chunk') {
+        return `块 ${activePreviewChunkIndex.value || ''}`.trim()
+      }
+      if (deleteDialogMode.value === 'file-chunks') {
+        return deleteDialogTarget.value?.filename || ''
+      }
+      return ''
+    })
+    
+    const deleteDialogDetail = computed(() => {
+      if (deleteDialogMode.value === 'chunk') {
+        return '这会移除当前选中的单个文本分片，适用于清理错误切分或无效内容。'
+      }
+      if (deleteDialogMode.value === 'file-chunks') {
+        return '这会清空当前文件已生成的全部分块，但不会删除文件本身。'
+      }
+      return ''
+    })
+    
+    const deleteDialogWarning = computed(() => {
+      if (deleteDialogMode.value === 'chunk') {
+        return '删除后不可恢复，和该分片关联的后续内容可能需要重新生成。'
+      }
+      if (deleteDialogMode.value === 'file-chunks') {
+        return '删除后不可恢复，该文件的全部分块将被清空，你需要重新执行分割才能恢复。'
+      }
+      return ''
+    })
+    
+    const deleteDialogConfirmText = computed(() => {
+      if (deleteDialogMode.value === 'chunk') return '确认删除分片'
+      if (deleteDialogMode.value === 'file-chunks') return '确认删除全部分块'
+      return '确认删除'
+    })
+    
+    const deleteDialogLoading = computed(() => {
+      if (deleteDialogMode.value === 'chunk') {
+        return !!deleteDialogTarget.value && deletingChunkId.value === deleteDialogTarget.value.id
+      }
+      if (deleteDialogMode.value === 'file-chunks') {
+        return !!deleteDialogTarget.value && deletingFileChunksId.value === deleteDialogTarget.value.id
+      }
+      return false
+    })
+    
+    const selectPreviewChunk = (chunkId) => {
+      selectedPreviewChunkId.value = chunkId
+    }
+    
+    const getChunkSnippet = (chunk) => {
+      return (chunk.editingContent || chunk.content || '')
+        .replace(/\s+/g, ' ')
+        .trim()
+        .slice(0, 110) || '空白分片'
+    }
+    
+    const jumpToChunk = () => {
+      const index = Number(previewJumpInput.value)
+      if (!Number.isInteger(index) || index < 1) {
+        ElMessage.warning('请输入有效的块号')
+        return
+      }
+    
+      const target = previewChunksWithIndex.value.find(chunk => chunk.displayIndex === index)
+      if (!target) {
+        ElMessage.warning('未找到对应块号')
+        return
+      }
+    
+      if (previewFilter.value === 'modified' && !isChunkModified(target)) {
+        previewFilter.value = 'all'
+      }
+    
+      selectedPreviewChunkId.value = target.id
+    }
+    
+    const resetChunk = (chunk) => {
+      chunk.editingContent = chunk.content
+    }
+    
+    const openDeleteChunkDialog = (chunk) => {
+      deleteDialogMode.value = 'chunk'
+      deleteDialogTarget.value = chunk
+      deleteDialogVisible.value = true
+    }
+    
+    const openDeleteFileChunksDialog = (file) => {
+      deleteDialogMode.value = 'file-chunks'
+      deleteDialogTarget.value = file
+      deleteDialogVisible.value = true
+    }
+    
+    const fetchPreviewChunks = async (fileId) => {
+      previewLoading.value = true
+      try {
+        const res = await chunkApi.list(projectId.value, { file_id: fileId })
+        previewChunks.value = (res.items || res || []).map(c => ({
+          ...c,
+          editingContent: c.content
+        }))
+        selectedPreviewChunkId.value = previewChunks.value[0]?.id || ''
+      } catch (e) {
+        console.error(e)
+        ElMessage.error('获取 chunks 失败')
+      } finally {
+        previewLoading.value = false
+      }
+    }
+    
+    const saveChunk = async (chunk) => {
+      savingChunks.value = true
+      try {
+        await chunkApi.update(projectId.value, chunk.id, {
+          content: chunk.editingContent
+        })
+        ElMessage.success('保存成功')
+        // Update local state
+        chunk.content = chunk.editingContent
+      } catch (e) {
+        console.error(e)
+        ElMessage.error('保存失败')
+      } finally {
+        savingChunks.value = false
+      }
+    }
+    
+    const deleteChunk = async (chunk) => {
+      deletingChunkId.value = chunk.id
+      try {
+        await chunkApi.delete(projectId.value, chunk.id)
+        previewChunks.value = previewChunks.value.filter(item => item.id !== chunk.id)
+    
+        if (previewFile.value?.id) {
+          const nextCount = Math.max((fileChunks.value[previewFile.value.id] || 1) - 1, 0)
+          if (nextCount > 0) {
+            fileChunks.value = {
+              ...fileChunks.value,
+              [previewFile.value.id]: nextCount
+            }
+          } else {
+            const { [previewFile.value.id]: _, ...rest } = fileChunks.value
+            fileChunks.value = rest
+          }
+        }
+    
+        ElMessage.success('删除成功')
+      } catch (e) {
+        console.error(e)
+        ElMessage.error('删除失败')
+      } finally {
+        deletingChunkId.value = ''
+        deleteDialogVisible.value = false
+        deleteDialogMode.value = ''
+        deleteDialogTarget.value = null
+      }
+    }
+    
+    const deleteFileChunks = async (file) => {
+      deletingFileChunksId.value = file.id
+      try {
+        const allChunks = []
+        let page = 1
+        let total = 0
+    
+        do {
+          const res = await chunkApi.list(projectId.value, {
+            file_id: file.id,
+            page,
+            page_size: 100
+          })
+          const items = res.items || []
+          total = res.total || items.length
+          allChunks.push(...items)
+          page += 1
+        } while (allChunks.length < total)
+    
+        for (const chunk of allChunks) {
+          await chunkApi.delete(projectId.value, chunk.id)
+        }
+    
+        const { [file.id]: _, ...rest } = fileChunks.value
+        fileChunks.value = rest
+    
+        if (previewFile.value?.id === file.id) {
+          previewChunks.value = []
+          selectedPreviewChunkId.value = ''
+        }
+    
+        ElMessage.success(`已删除 ${allChunks.length} 个分块`)
+      } catch (e) {
+        console.error(e)
+        ElMessage.error('删除全部分块失败')
+      } finally {
+        deletingFileChunksId.value = ''
+        deleteDialogVisible.value = false
+        deleteDialogMode.value = ''
+        deleteDialogTarget.value = null
+      }
+    }
+    
+    const confirmDeleteAction = async () => {
+      if (!deleteDialogTarget.value) return
+    
+      if (deleteDialogMode.value === 'chunk') {
+        await deleteChunk(deleteDialogTarget.value)
+        return
+      }
+    
+      if (deleteDialogMode.value === 'file-chunks') {
+        await deleteFileChunks(deleteDialogTarget.value)
+      }
+    }
+    
+    watch(filteredPreviewChunks, (chunks) => {
+      if (!chunks.length) {
+        selectedPreviewChunkId.value = ''
+        return
+      }
+    
+      if (!chunks.some(chunk => chunk.id === selectedPreviewChunkId.value)) {
+        selectedPreviewChunkId.value = chunks[0].id
+      }
+    })
+    
+    const refreshFiles = () => {
+      fetchFiles()
+    }
+    
+    const startFilePolling = () => {
+      if (filePollingTimer.value) return
+      filePollingTimer.value = window.setInterval(() => {
+        fetchFiles()
+      }, 3000)
+    }
+    
+    const stopFilePolling = () => {
+      if (!filePollingTimer.value) return
+      window.clearInterval(filePollingTimer.value)
+      filePollingTimer.value = null
+    }
+    
+    const formatSize = (bytes) => {
+      if (!bytes) return '0 B'
+      const units = ['B', 'KB', 'MB', 'GB']
+      let i = 0
+      while (bytes >= 1024 && i < units.length - 1) {
+        bytes /= 1024
+        i++
+      }
+      return `${bytes.toFixed(1)} ${units[i]}`
+    }
+    
+    const getFileBg = (type) => {
+      const colors = {
+        pdf: '#ef4444',
+        docx: '#3b82f6',
+        xlsx: '#22c55e',
+        csv: '#f59e0b',
+        md: '#8b5cf6',
+        txt: '#6b7280',
+        epub: '#ec4899'
+      }
+      return colors[type] || '#6b7280'
+    }
+    
+    const getFileIcon = (type) => {
+      const icons = {
+        pdf: 'Document',
+        docx: 'Document',
+        xlsx: 'Grid',
+        csv: 'Grid',
+        md: 'Document',
+        txt: 'Document',
+        epub: 'Book'
+      }
+      return icons[type] || 'Document'
+    }
+    
+    onMounted(() => {
+      fetchAvailableModels()
+      fetchFiles()
+    })
+    
+    watch(processingCount, (count) => {
+      if (count > 0) {
+        startFilePolling()
+      } else {
+        stopFilePolling()
+      }
+    }, { immediate: true })
+    
+    onUnmounted(() => {
+      stopFilePolling()
+    })
+
+    return {
+      router,
+      route,
+      projectId,
+      loading,
+      splitting,
+      files,
+      filterStatus,
+      fileChunks,
+      isInitialLoad,
+      availableModels,
+      filePollingTimer,
+      generateDialogVisible,
+      generatingQuestions,
+      DEFAULT_GENERATE_PROMPT,
+      selectedFiles,
+      splitDialogVisible,
+      chunkPreviewVisible,
+      previewFile,
+      previewChunks,
+      previewLoading,
+      savingChunks,
+      deletingChunkId,
+      deletingFileChunksId,
+      deleteDialogVisible,
+      deleteDialogMode,
+      deleteDialogTarget,
+      previewSearch,
+      previewFilter,
+      previewJumpInput,
+      selectedPreviewChunkId,
+      isAllSelected,
+      selectedCount,
+      filteredFiles,
+      toggleSelectAll,
+      toggleSelect,
+      isSelected,
+      clearSelection,
+      splitConfig,
+      generateConfig,
+      methods,
+      completedFiles,
+      processingCount,
+      totalChunks,
+      normalizeModelType,
+      embeddingModels,
+      selectedEmbeddingModel,
+      generateModels,
+      getProviderLabel,
+      fetchAllChunks,
+      goToModelSettings,
+      fetchAvailableModels,
+      fetchFiles,
+      fetchChunksCount,
+      openSplitDialog,
+      handleBatchSplit,
+      openGenerateDialog,
+      handleBatchGenerate,
+      openChunkPreview,
+      previewChunksWithIndex,
+      isChunkModified,
+      filteredPreviewChunks,
+      activePreviewChunk,
+      activePreviewChunkIndex,
+      modifiedPreviewCount,
+      deleteDialogTitle,
+      deleteDialogItemName,
+      deleteDialogDetail,
+      deleteDialogWarning,
+      deleteDialogConfirmText,
+      deleteDialogLoading,
+      selectPreviewChunk,
+      getChunkSnippet,
+      jumpToChunk,
+      resetChunk,
+      openDeleteChunkDialog,
+      openDeleteFileChunksDialog,
+      fetchPreviewChunks,
+      saveChunk,
+      deleteChunk,
+      deleteFileChunks,
+      confirmDeleteAction,
+      refreshFiles,
+      startFilePolling,
+      stopFilePolling,
+      formatSize,
+      getFileBg,
+      getFileIcon
+    }
+  }
+})

frontend/src/pages/CrawlerPage.vue

@@ -0,0 +1,386 @@
+<template>
+  <div class="crawler-page">
+    <div class="page-header">
+      <div class="header-content">
+        <div class="header-icon">
+          <svg width="32" height="32" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+            <circle cx="12" cy="12" r="10"/>
+            <path d="M2 12h20M12 2a15.3 15.3 0 0 1 4 10 15.3 15.3 0 0 1-4 10 15.3 15.3 0 0 1-4-10 15.3 15.3 0 0 1 4-10z"/>
+          </svg>
+        </div>
+        <div class="header-text">
+          <h1>数据爬虫</h1>
+          <p>从网页自动采集数据，用于构建训练数据集</p>
+        </div>
+      </div>
+    </div>
+
+    <div class="crawler-content">
+      <!-- Crawler Config Card -->
+      <div class="config-card">
+        <h2>爬取配置</h2>
+
+        <el-form :model="form" label-position="top">
+          <el-form-item label="目标网址">
+            <el-input
+              v-model="form.url"
+              placeholder="https://example.com"
+              :prefix-icon="Link"
+            >
+              <template #prepend>
+                <el-select v-model="form.method" style="width: 100px">
+                  <el-option label="GET" value="GET" />
+                  <el-option label="POST" value="POST" />
+                </el-select>
+              </template>
+            </el-input>
+          </el-form-item>
+
+          <el-form-item label="选择项目">
+            <el-select v-model="form.projectId" placeholder="选择目标项目" style="width: 100%">
+              <el-option
+                v-for="project in projects"
+                :key="project.id"
+                :label="project.name"
+                :value="project.id"
+              />
+            </el-select>
+          </el-form-item>
+
+          <el-form-item label="爬取规则">
+            <div class="rule-options">
+              <el-checkbox v-model="form.extractTitle">提取标题</el-checkbox>
+              <el-checkbox v-model="form.extractContent">提取正文内容</el-checkbox>
+              <el-checkbox v-model="form.extractLinks">提取所有链接</el-checkbox>
+              <el-checkbox v-model="form.extractImages">提取图片链接</el-checkbox>
+            </div>
+          </el-form-item>
+
+          <el-form-item label="CSS 选择器 (可选)">
+            <el-input
+              v-model="form.cssSelector"
+              placeholder="如: article.content, .post-body"
+            />
+          </el-form-item>
+
+          <el-form-item label="爬取深度">
+            <el-slider v-model="form.depth" :min="1" :max="5" show-input />
+          </el-form-item>
+
+          <el-form-item>
+            <el-button
+              type="primary"
+              :loading="crawling"
+              @click="startCrawl"
+              class="start-btn"
+            >
+              <el-icon><Crawler /></el-icon>
+              {{ crawling ? '爬取中...' : '开始爬取' }}
+            </el-button>
+          </el-form-item>
+        </el-form>
+      </div>
+
+      <!-- Results Card -->
+      <div class="results-card">
+        <div class="results-header">
+          <h2>爬取结果</h2>
+          <span class="result-count" v-if="results.length">{{ results.length }} 条</span>
+        </div>
+
+        <div class="results-content" v-loading="crawling">
+          <div v-if="!crawling && results.length === 0" class="empty-results">
+            <el-icon class="empty-icon"><Link /></el-icon>
+            <p>配置完成后点击"开始爬取"</p>
+          </div>
+
+          <div v-else class="results-list">
+            <div
+              v-for="(item, index) in results"
+              :key="index"
+              class="result-item"
+            >
+              <div class="result-title">{{ item.title || '无标题' }}</div>
+              <div class="result-url">{{ item.url }}</div>
+              <div class="result-preview" v-if="item.content">
+                {{ item.content.substring(0, 150) }}...
+              </div>
+              <div class="result-meta">
+                <el-tag size="small" v-if="item.images?.length">
+                  {{ item.images.length }} 张图片
+                </el-tag>
+                <el-tag size="small" v-if="item.links?.length">
+                  {{ item.links.length }} 个链接
+                </el-tag>
+              </div>
+            </div>
+          </div>
+        </div>
+
+        <div class="results-actions" v-if="results.length > 0">
+          <el-button @click="exportResults">
+            <el-icon><Download /></el-icon>
+            导出数据
+          </el-button>
+          <el-button type="primary" @click="saveToProject">
+            <el-icon><FolderAdd /></el-icon>
+            保存到项目
+          </el-button>
+        </div>
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { ref, onMounted } from 'vue'
+import { useRouter } from 'vue-router'
+import { ElMessage } from 'element-plus'
+import { Link, Download, FolderAdd } from '@element-plus/icons-vue'
+import { projectApi } from '@/core/api'
+
+const router = useRouter()
+
+const projects = ref([])
+const crawling = ref(false)
+const results = ref([])
+
+const form = ref({
+  url: '',
+  method: 'GET',
+  projectId: '',
+  extractTitle: true,
+  extractContent: true,
+  extractLinks: false,
+  extractImages: false,
+  cssSelector: '',
+  depth: 1
+})
+
+const fetchProjects = async () => {
+  try {
+    const res = await projectApi.list()
+    projects.value = res.items || res || []
+  } catch (error) {
+    projects.value = []
+  }
+}
+
+const startCrawl = async () => {
+  if (!form.value.url) {
+    ElMessage.warning('请输入目标网址')
+    return
+  }
+
+  if (!form.value.projectId) {
+    ElMessage.warning('请选择目标项目')
+    return
+  }
+
+  crawling.value = true
+  results.value = []
+
+  try {
+    // Simulate crawling - in production this would call the backend API
+    await new Promise(resolve => setTimeout(resolve, 2000))
+
+    // Demo results
+    results.value = [
+      {
+        title: '示例页面标题',
+        url: form.value.url,
+        content: '这是从网页中提取的内容示例。爬虫会解析HTML结构，提取文本、图片链接和其他有价值的数据。',
+        images: ['https://example.com/image1.jpg'],
+        links: ['https://example.com/page1', 'https://example.com/page2']
+      },
+      {
+        title: '子页面标题 1',
+        url: form.value.url + '/page1',
+        content: '这是子页面的内容...',
+        images: [],
+        links: []
+      }
+    ]
+
+    ElMessage.success('爬取完成')
+  } catch (error) {
+    ElMessage.error('爬取失败: ' + error.message)
+  } finally {
+    crawling.value = false
+  }
+}
+
+const exportResults = () => {
+  const data = JSON.stringify(results.value, null, 2)
+  const blob = new Blob([data], { type: 'application/json' })
+  const url = URL.createObjectURL(blob)
+  const a = document.createElement('a')
+  a.href = url
+  a.download = 'crawler-results.json'
+  a.click()
+  URL.revokeObjectURL(url)
+}
+
+const saveToProject = () => {
+  ElMessage.success('数据已保存到项目')
+}
+
+onMounted(() => fetchProjects())
+</script>
+
+<style scoped>
+.crawler-page {
+  min-height: 100vh;
+  padding: 40px;
+  max-width: 1200px;
+  margin: 0 auto;
+}
+
+.page-header {
+  margin-bottom: 40px;
+}
+
+.header-content {
+  display: flex;
+  align-items: center;
+  gap: 20px;
+}
+
+.header-icon {
+  width: 64px;
+  height: 64px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: var(--accent-primary-muted);
+  border-radius: var(--radius-lg);
+  color: var(--accent-primary);
+}
+
+.header-text h1 {
+  font-size: 28px;
+  font-weight: 600;
+  margin-bottom: 4px;
+}
+
+.header-text p {
+  color: var(--text-secondary);
+}
+
+.crawler-content {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  gap: 24px;
+}
+
+.config-card,
+.results-card {
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-subtle);
+  border-radius: var(--radius-lg);
+  padding: 24px;
+}
+
+.config-card h2,
+.results-card h2 {
+  font-size: 18px;
+  font-weight: 600;
+  margin-bottom: 20px;
+}
+
+.rule-options {
+  display: flex;
+  flex-direction: column;
+  gap: 8px;
+}
+
+.start-btn {
+  width: 100%;
+  padding: 14px;
+  font-size: 15px;
+}
+
+.results-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  margin-bottom: 16px;
+}
+
+.result-count {
+  font-size: 14px;
+  color: var(--text-secondary);
+  background: var(--accent-primary-muted);
+  padding: 4px 12px;
+  border-radius: 100px;
+}
+
+.results-content {
+  min-height: 300px;
+}
+
+.empty-results {
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  justify-content: center;
+  height: 300px;
+  color: var(--text-tertiary);
+}
+
+.empty-icon {
+  font-size: 48px;
+  margin-bottom: 16px;
+  opacity: 0.3;
+}
+
+.results-list {
+  display: flex;
+  flex-direction: column;
+  gap: 12px;
+}
+
+.result-item {
+  padding: 16px;
+  background: var(--bg-tertiary);
+  border-radius: var(--radius-md);
+  border: 1px solid var(--border-subtle);
+}
+
+.result-title {
+  font-weight: 600;
+  margin-bottom: 4px;
+}
+
+.result-url {
+  font-size: 12px;
+  color: var(--accent-primary);
+  margin-bottom: 8px;
+}
+
+.result-preview {
+  font-size: 13px;
+  color: var(--text-secondary);
+  margin-bottom: 8px;
+  line-height: 1.5;
+}
+
+.result-meta {
+  display: flex;
+  gap: 8px;
+}
+
+.results-actions {
+  display: flex;
+  gap: 12px;
+  margin-top: 16px;
+  padding-top: 16px;
+  border-top: 1px solid var(--border-subtle);
+}
+
+@media (max-width: 900px) {
+  .crawler-content {
+    grid-template-columns: 1fr;
+  }
+}
+</style>

frontend/src/pages/HomePage.vue

@@ -0,0 +1,390 @@
+<template>
+  <div class="home">
+    <!-- Hero Section -->
+    <section class="hero">
+      <div class="hero-content">
+        <!-- Logo -->
+        <div class="hero-logo">
+          <svg width="56" height="56" viewBox="0 0 56 56" fill="none" xmlns="http://www.w3.org/2000/svg">
+            <defs>
+              <linearGradient id="logoGradient" x1="0%" y1="0%" x2="100%" y2="100%">
+                <stop offset="0%" style="stop-color:#00d4ff"/>
+                <stop offset="100%" style="stop-color:#7c3aed"/>
+              </linearGradient>
+            </defs>
+            <!-- 外圈 - 数据集合 -->
+            <rect x="4" y="4" width="48" height="48" rx="12" stroke="url(#logoGradient)" stroke-width="2.5" fill="none" opacity="0.3"/>
+            <!-- Y 字母 - 数据流/分支 -->
+            <path d="M18 42V22L28 12V18" stroke="url(#logoGradient)" stroke-width="3.5" stroke-linecap="round" stroke-linejoin="round" fill="none"/>
+            <path d="M28 18L38 28" stroke="url(#logoGradient)" stroke-width="3.5" stroke-linecap="round" fill="none"/>
+            <!-- 数据节点 - 神经网络样式 -->
+            <circle cx="18" cy="42" r="3" fill="#00d4ff"/>
+            <circle cx="28" cy="12" r="3" fill="#7c3aed"/>
+            <circle cx="38" cy="28" r="3" fill="#00d4ff"/>
+            <circle cx="28" cy="18" r="2.5" fill="#00d4ff" opacity="0.7"/>
+            <!-- 连接线 - 数据流向 -->
+            <circle cx="28" cy="32" r="2" fill="#7c3aed" opacity="0.5"/>
+            <circle cx="20" cy="32" r="1.5" fill="#00d4ff" opacity="0.4"/>
+            <circle cx="36" cy="38" r="1.5" fill="#7c3aed" opacity="0.4"/>
+          </svg>
+          <span class="logo-text">YG<span class="logo-highlight">Datasets</span></span>
+        </div>
+
+        <div class="hero-badge">
+          <span class="badge-dot"></span>
+          <span>AI 驱动数据生成</span>
+        </div>
+        <h1 class="hero-title">
+          构建高质量<br />
+          <span class="glow-text">训练数据集</span>
+        </h1>
+        <p class="hero-subtitle">
+          通过智能分割、AI 生成问答和无缝评估，
+          将文档转化为结构化数据集。
+        </p>
+        <div class="hero-actions">
+          <el-button type="primary" size="large" @click="createProject" class="btn-primary">
+            <el-icon><Plus /></el-icon>
+            创建项目
+          </el-button>
+          <el-button size="large" @click="goToCrawler" class="btn-secondary">
+            <el-icon><Connection /></el-icon>
+            数据爬虫
+          </el-button>
+          <el-button size="large" @click="goToModels" class="btn-secondary">
+            <el-icon><Cpu /></el-icon>
+            模型管理
+          </el-button>
+        </div>
+      </div>
+
+      <!-- Hero Visual - Modern Abstract Composition -->
+      <div class="hero-visual">
+        <!-- Galaxy Background -->
+        <div class="galaxy-bg">
+          <!-- Nebula clouds -->
+          <div class="nebula-cloud nebula-1"></div>
+          <div class="nebula-cloud nebula-2"></div>
+          <div class="nebula-cloud nebula-3"></div>
+
+          <!-- Galaxy core -->
+          <div class="galaxy-core"></div>
+
+          <!-- Spiral arms -->
+          <div class="galaxy-spiral">
+            <div class="spiral-arm spiral-arm-1"></div>
+            <div class="spiral-arm spiral-arm-2"></div>
+            <div class="spiral-arm spiral-arm-3"></div>
+          </div>
+
+          <!-- Orbit rings with stars -->
+          <div class="orbit-ring orbit-ring-1">
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+          </div>
+          <div class="orbit-ring orbit-ring-2">
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+          </div>
+          <div class="orbit-ring orbit-ring-3">
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+          </div>
+          <div class="orbit-ring orbit-ring-4">
+            <span class="orbit-star"></span>
+            <span class="orbit-star"></span>
+          </div>
+        </div>
+
+        <!-- Light rays -->
+        <div class="light-rays">
+          <div class="ray"></div>
+          <div class="ray"></div>
+          <div class="ray"></div>
+          <div class="ray"></div>
+          <div class="ray"></div>
+        </div>
+
+        <!-- Ambient particles -->
+        <span class="ambient-particle"></span>
+        <span class="ambient-particle"></span>
+        <span class="ambient-particle"></span>
+        <span class="ambient-particle"></span>
+        <span class="ambient-particle"></span>
+
+        <!-- Abstract background orbs -->
+        <div class="orb orb-1"></div>
+        <div class="orb orb-2"></div>
+        <div class="orb orb-3"></div>
+
+        <!-- Central floating UI element -->
+        <div class="floating-ui">
+          <div class="ui-header">
+            <div class="ui-dot"></div>
+            <div class="ui-dot"></div>
+            <div class="ui-dot"></div>
+          </div>
+          <div class="ui-content">
+            <div class="ui-line"></div>
+            <div class="ui-line short"></div>
+            <div class="ui-line"></div>
+          </div>
+          <div class="ui-badge">
+            <el-icon><Check /></el-icon>
+            <span>处理完成</span>
+          </div>
+        </div>
+
+        <!-- Floating feature pills - main features -->
+        <div class="feature-pill pill-1">
+          <el-icon><Document /></el-icon>
+          <span>多格式支持</span>
+        </div>
+        <div class="feature-pill pill-2">
+          <el-icon><MagicStick /></el-icon>
+          <span>AI 生成</span>
+        </div>
+        <div class="feature-pill pill-3">
+          <el-icon><DataAnalysis /></el-icon>
+          <span>智能评估</span>
+        </div>
+
+        <!-- Additional floating labels -->
+        <div class="feature-pill pill-4">
+          <el-icon><Connection /></el-icon>
+          <span>API 集成</span>
+        </div>
+        <div class="feature-pill pill-5">
+          <el-icon><Clock /></el-icon>
+          <span>批量处理</span>
+        </div>
+        <div class="feature-pill pill-6">
+          <el-icon><Lock /></el-icon>
+          <span>数据安全</span>
+        </div>
+        <div class="feature-pill pill-7">
+          <el-icon><TrendCharts /></el-icon>
+          <span>可视化</span>
+        </div>
+      </div>
+    </section>
+
+    <!-- Projects Section -->
+    <section class="projects-section">
+      <div class="section-header">
+        <div class="section-title">
+          <h2>我的项目</h2>
+          <p>{{ total }} 个项目</p>
+        </div>
+        <el-button type="primary" @click="createProject" class="add-btn">
+          <el-icon><Plus /></el-icon>
+          新建
+        </el-button>
+      </div>
+
+      <!-- Projects Grid -->
+      <div class="projects-grid" v-loading="loading">
+        <!-- Empty State -->
+        <EmptyState
+          v-if="!loading && projects.length === 0"
+          :icon="FolderAdd"
+          title="暂无项目"
+          description="创建您的第一个项目开始生成数据集"
+          action-text="创建项目"
+          @action="createProject"
+        />
+
+        <!-- Project Cards -->
+        <ProjectCard
+          v-else
+          v-for="(project, index) in projects"
+          :key="project.id"
+          :project="project"
+          :index="index"
+          @click="openProject"
+          @delete="confirmDelete"
+        />
+      </div>
+
+      <!-- Pagination -->
+      <div class="pagination-wrapper" v-if="needPagination">
+        <div class="pagination-minimal">
+          <span class="page-info">第 {{ currentPage }} / {{ totalPages }} 页</span>
+          <div class="page-arrows">
+            <button
+              class="arrow-btn"
+              :disabled="currentPage === 1"
+              @click="handlePageChange(currentPage - 1)"
+            >
+              <el-icon><ArrowLeft /></el-icon>
+            </button>
+            <button
+              class="arrow-btn"
+              :disabled="currentPage === totalPages"
+              @click="handlePageChange(currentPage + 1)"
+            >
+              <el-icon><ArrowRight /></el-icon>
+            </button>
+          </div>
+        </div>
+      </div>
+    </section>
+
+    <!-- Create Dialog -->
+    <CreateProjectDialog
+      v-model:visible="dialogVisible"
+      :loading="submitting"
+      @submit="handleCreateSubmit"
+    />
+
+    <!-- Delete Confirmation Dialog -->
+    <DeleteDialog
+      v-model:visible="deleteDialogVisible"
+      :item-name="projectToDelete?.name"
+      :loading="deleting"
+      @confirm="handleDelete"
+    />
+  </div>
+</template>
+
+<script setup lang="ts">
+import { ref, onMounted, computed } from 'vue'
+import { useRouter } from 'vue-router'
+import { ElMessage } from 'element-plus'
+import { FolderAdd, Check, Connection, Clock, Lock, TrendCharts, ArrowLeft, ArrowRight } from '@element-plus/icons-vue'
+import { projectApi } from '@/core/api'
+import type { Project, ProjectCreate } from '@/shared/types'
+
+// Components
+import EmptyState from '@/shared/components/common/EmptyState.vue'
+import ProjectCard from '@/shared/components/common/ProjectCard.vue'
+import CreateProjectDialog from '@/shared/components/common/CreateProjectDialog.vue'
+import DeleteDialog from '@/shared/components/common/DeleteDialog.vue'
+
+const router = useRouter()
+
+const loading = ref(false)
+const projects = ref([])
+const dialogVisible = ref(false)
+const deleteDialogVisible = ref(false)
+const projectToDelete = ref(null)
+const submitting = ref(false)
+const deleting = ref(false)
+
+// Pagination
+const currentPage = ref(1)
+const pageSize = ref(9)
+const total = ref(0)
+
+const fetchProjects = async () => {
+  loading.value = true
+  try {
+    const res = await projectApi.list({ page: currentPage.value, page_size: pageSize.value })
+    // API returns: { items: [], total, page, page_size, total_pages }
+    if (res && typeof res === 'object' && 'items' in res) {
+      projects.value = res.items || []
+      total.value = res.total || 0
+    } else if (Array.isArray(res)) {
+      projects.value = res
+      total.value = res.length
+    } else {
+      projects.value = []
+      total.value = 0
+    }
+  } catch (error) {
+    projects.value = []
+    total.value = 0
+  } finally {
+    loading.value = false
+  }
+}
+
+const handlePageChange = (page: number) => {
+  currentPage.value = page
+  fetchProjects()
+}
+
+const needPagination = computed(() => total.value > pageSize.value || projects.value.length === pageSize.value)
+
+const totalPages = computed(() => Math.ceil(total.value / pageSize.value))
+
+
+const createProject = () => {
+  dialogVisible.value = true
+}
+
+const handleCreateSubmit = async (formData) => {
+  // Validation - name, description and type are required
+  if (!formData.name || formData.name.trim() === '') {
+    ElMessage.warning('请输入项目名称')
+    return
+  }
+  if (!formData.description || formData.description.trim() === '') {
+    ElMessage.warning('请输入项目描述')
+    return
+  }
+  if (!formData.type) {
+    ElMessage.warning('请选择项目类型')
+    return
+  }
+
+  console.log('Creating project with form:', formData)
+  submitting.value = true
+  try {
+    const res = await projectApi.create(formData)
+    console.log('Create response:', res)
+    ElMessage.success('项目创建成功')
+    dialogVisible.value = false
+    fetchProjects()
+    // New format: {id: "..."}
+    const projectId = res.id
+    console.log('Navigating to:', projectId)
+    router.push(`/project/${projectId}`)
+  } catch (error) {
+    console.error('Create project error:', error)
+    ElMessage.error('创建项目失败: ' + (error.message || '未知错误'))
+  } finally {
+    submitting.value = false
+  }
+}
+
+const openProject = (project) => {
+  router.push(`/project/${project.id}`)
+}
+
+const confirmDelete = (project) => {
+  projectToDelete.value = project
+  deleteDialogVisible.value = true
+}
+
+const handleDelete = async () => {
+  if (!projectToDelete.value) return
+  deleting.value = true
+  try {
+    await projectApi.delete(projectToDelete.value.id)
+    ElMessage.success('项目已删除')
+    deleteDialogVisible.value = false
+    projectToDelete.value = null
+    fetchProjects()
+  } catch (error) {
+    ElMessage.error('删除失败')
+  } finally {
+    deleting.value = false
+  }
+}
+
+const goToDataSquare = () => router.push('/data-square')
+const goToCrawler = () => router.push('/crawler')
+const goToModels = () => router.push('/models')
+
+onMounted(() => fetchProjects())
+</script>
+
+<style scoped>
+@import '@/styles/pages/home.scss';
+</style>

frontend/src/pages/ModelSettingsPage.vue

@@ -0,0 +1,262 @@
+<template>
+  <div class="model-settings">
+    <!-- 背景效果 -->
+    <div class="bg-effects">
+      <div class="glow-orb glow-1"></div>
+      <div class="glow-orb glow-2"></div>
+    </div>
+
+    <!-- 页面头部 -->
+    <header class="page-header">
+      <div class="header-left">
+        <el-button text class="back-btn" @click="goHome">
+          <el-icon><ArrowLeft /></el-icon>
+          <span>返回</span>
+        </el-button>
+      </div>
+      <div class="header-content">
+        <h1 class="page-title">
+          <el-icon class="title-icon"><Cpu /></el-icon>
+          模型管理
+        </h1>
+        <p class="page-subtitle">管理您的 AI 模型 API</p>
+      </div>
+      <div class="header-right">
+        <el-button type="primary" class="add-btn" @click="openAddDialog">
+          <el-icon><Plus /></el-icon>
+          <span>添加模型</span>
+        </el-button>
+      </div>
+    </header>
+
+    <!-- 主内容 -->
+    <main class="page-main">
+      <!-- 模型列表 -->
+      <section class="models-section">
+        <div class="section-header">
+          <h2 class="section-title">
+            <span class="title-line"></span>
+            已配置的模型
+          </h2>
+          <span class="count-badge">{{ models.length }} 个</span>
+        </div>
+
+        <!-- 空状态 -->
+        <div v-if="models.length === 0 && !loading" class="empty-state">
+          <div class="empty-illustration">
+            <div class="pulse-ring"></div>
+            <el-icon size="48"><Setting /></el-icon>
+          </div>
+          <h3>暂无模型配置</h3>
+          <p>添加您的第一个 AI 模型开始使用</p>
+          <el-button type="primary" @click="openAddDialog">添加模型</el-button>
+        </div>
+
+        <!-- 模型卡片 -->
+        <div v-else class="models-grid">
+          <article
+            v-for="(model, index) in models"
+            :key="model.id"
+            class="model-card"
+            :class="{ 'is-default': model.is_default === 'true' }"
+            :style="{ '--delay': index * 0.08 + 's' }"
+          >
+            <div class="card-glow"></div>
+
+            <!-- 默认标识 -->
+            <div v-if="model.is_default === 'true'" class="default-badge">
+              <el-icon><Star /></el-icon>
+              默认
+            </div>
+
+            <!-- 提供商图标 -->
+            <div class="provider-logo" :class="model.provider">
+              {{ getProviderAbbr(model.provider) }}
+            </div>
+
+            <!-- 模型信息 -->
+            <div class="model-info">
+              <div class="model-name-row">
+                <h3 class="model-name">{{ model.model_name }}</h3>
+                <span class="model-type-badge" :class="`type-${normalizeModelType(model.model_type, model.model_name)}`">
+                  {{ getModelTypeLabel(model.model_type, model.model_name) }}
+                </span>
+              </div>
+              <p class="model-endpoint">
+                <el-icon><Link /></el-icon>
+                {{ model.api_base || '默认端点' }}
+              </p>
+            </div>
+
+            <!-- 底部操作 -->
+            <div class="card-footer">
+              <div class="status-badge" :class="model.connection_status">
+                <span class="status-dot" :class="model.connection_status"></span>
+                <template v-if="model.connection_status === 'connected'">已联通</template>
+                <template v-else-if="model.connection_status === 'failed'">连接失败</template>
+                <template v-else>未测试</template>
+              </div>
+              <div class="card-actions">
+                <el-button text class="action-btn test" @click="testConnection(model)">
+                  测试连接
+                </el-button>
+                <el-button text class="action-btn delete" @click="confirmDelete(model)">
+                  删除
+                </el-button>
+              </div>
+            </div>
+          </article>
+        </div>
+      </section>
+    </main>
+
+    <el-dialog
+      v-model="showAddDialog"
+      :show-close="false"
+      width="560px"
+      class="add-dialog"
+      :append-to-body="true"
+    >
+      <template #header>
+        <div class="dialog-header">
+          <div class="dialog-icon">
+            <el-icon size="20"><Plus /></el-icon>
+          </div>
+          <div class="dialog-title">
+            <h3>添加模型</h3>
+            <p>配置新的 AI 模型</p>
+          </div>
+          <button class="dialog-close" @click="showAddDialog = false">
+            <el-icon><Close /></el-icon>
+          </button>
+        </div>
+      </template>
+
+      <el-form :model="modelForm" label-position="top" class="model-form">
+        <el-form-item label="选择提供商">
+          <el-select
+            v-model="modelForm.provider"
+            placeholder="选择 AI 服务提供商"
+            size="large"
+            class="provider-select"
+            popper-class="provider-select-dropdown"
+          >
+            <el-option
+              v-for="provider in providers"
+              :key="provider.value"
+              :label="provider.label"
+              :value="provider.value"
+            >
+              <div class="provider-option-item">
+                <span class="provider-icon">{{ provider.abbr }}</span>
+                <div class="provider-copy">
+                  <span>{{ provider.label }}</span>
+                  <small>{{ provider.desc }}</small>
+                </div>
+              </div>
+            </el-option>
+          </el-select>
+        </el-form-item>
+
+        <el-form-item label="模型类型">
+          <el-select
+            v-model="modelForm.model_type"
+            placeholder="选择模型类型"
+            size="large"
+            class="provider-select"
+          >
+            <el-option
+              v-for="type in modelTypes"
+              :key="type.value"
+              :label="type.label"
+              :value="type.value"
+            >
+              <div class="provider-option-item">
+                <span class="provider-icon">{{ type.abbr }}</span>
+                <div class="provider-copy">
+                  <span>{{ type.label }}</span>
+                  <small>{{ type.desc }}</small>
+                </div>
+              </div>
+            </el-option>
+          </el-select>
+        </el-form-item>
+
+        <el-form-item label="模型名称" required>
+          <el-input
+            v-model="modelForm.model_name"
+            placeholder="例如: gpt-4o-mini / text-embedding-v3-small"
+            size="large"
+          />
+        </el-form-item>
+
+        <el-form-item label="API Key" required>
+          <el-input
+            v-model="modelForm.api_key"
+            type="password"
+            placeholder="输入 API Key"
+            size="large"
+            show-password
+          />
+        </el-form-item>
+
+        <el-form-item label="API 地址">
+          <el-input
+            v-model="modelForm.api_base"
+            placeholder="自定义 API 地址"
+            size="large"
+          />
+        </el-form-item>
+
+        <el-form-item>
+          <el-checkbox v-model="modelForm.is_default">
+            设为默认模型
+          </el-checkbox>
+        </el-form-item>
+      </el-form>
+
+      <template #footer>
+        <div class="dialog-footer">
+          <el-button @click="showAddDialog = false" size="large">取消</el-button>
+          <el-button type="primary" @click="addModel" :loading="submitting" size="large">
+            添加模型
+          </el-button>
+        </div>
+      </template>
+    </el-dialog>
+
+    <el-dialog
+      v-model="deleteDialogVisible"
+      :show-close="false"
+      width="400"
+      class="delete-dialog"
+      :append-to-body="true"
+    >
+      <template #header>
+        <div class="delete-header">
+          <div class="delete-icon">
+            <el-icon size="24"><WarningFilled /></el-icon>
+          </div>
+          <h3>确认删除</h3>
+        </div>
+      </template>
+
+      <div class="delete-content">
+        <p>确定要删除模型 <strong>{{ modelToDelete?.model_name }}</strong> 吗？</p>
+        <p class="warning-text">此操作不可恢复</p>
+      </div>
+
+      <template #footer>
+        <div class="delete-footer">
+          <el-button @click="deleteDialogVisible = false" size="large">取消</el-button>
+          <el-button type="danger" @click="handleDelete" :loading="deleting" size="large">
+            确认删除
+          </el-button>
+        </div>
+      </template>
+    </el-dialog>
+  </div>
+</template>
+
+<script lang="ts" src="../page-logic/ModelSettingsPage.ts"></script>
+<style scoped src="../styles/pages/model-settings.css"></style>

frontend/src/pages/ProjectDatasetPage.vue

@@ -128,7 +128,7 @@
 import { ref, reactive, computed, onMounted } from 'vue'
 import { useRoute } from 'vue-router'
 import { ElMessage, ElMessageBox } from 'element-plus'
-import { datasetApi } from '@/api'
+import { datasetApi } from '@/core/api'
 
 const route = useRoute()
 const projectId = computed(() => route.params.id)

frontend/src/pages/ProjectFilePage.vue

@@ -0,0 +1,319 @@
+<template>
+  <div class="file-manage">
+    <!-- Header -->
+    <div class="page-header">
+      <div class="header-left">
+        <h2 class="page-title">文件管理</h2>
+        <p class="page-subtitle">管理您的文档集合</p>
+      </div>
+      <div class="header-actions">
+        <el-button type="primary" @click="handleUpload" class="upload-btn">
+          <el-icon><Upload /></el-icon>
+          <span>上传文件</span>
+        </el-button>
+      </div>
+    </div>
+
+    <!-- Stats Cards -->
+    <div class="stats-grid">
+      <div
+        class="stat-card stat-total"
+        :class="{ active: filterStatus === '' }"
+        @click="filterStatus = ''"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><Document /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ files.length }}</span>
+            <span class="stat-label">总文件数</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-completed"
+        :class="{ active: filterStatus === 'completed' }"
+        @click="filterStatus = filterStatus === 'completed' ? '' : 'completed'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><CircleCheckFilled /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ completedFiles }}</span>
+            <span class="stat-label">已完成</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-processing"
+        :class="{ active: filterStatus === 'processing' }"
+        @click="filterStatus = filterStatus === 'processing' ? '' : 'processing'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><Loading /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ processingFiles.length }}</span>
+            <span class="stat-label">处理中</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-failed"
+        :class="{ active: filterStatus === 'failed' }"
+        @click="filterStatus = filterStatus === 'failed' ? '' : 'failed'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><CircleCloseFilled /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ failedFiles }}</span>
+            <span class="stat-label">失败</span>
+          </div>
+        </div>
+      </div>
+    </div>
+
+    <!-- File List Container -->
+    <div class="file-container" v-loading="loading && isInitialLoad">
+      <!-- Empty State -->
+      <div v-if="!loading && !isInitialLoad && filteredFiles.length === 0 && !uploading" class="empty-state">
+        <div class="empty-illustration">
+          <div class="orbit orbit-1"></div>
+          <div class="orbit orbit-2"></div>
+          <div class="orbit orbit-3"></div>
+          <div class="empty-core">
+            <el-icon size="40"><FolderOpened /></el-icon>
+          </div>
+        </div>
+        <h3 class="empty-title">暂无文件</h3>
+        <p class="empty-desc">上传您的第一个文档，开启智能处理之旅</p>
+      </div>
+
+      <!-- Files Table -->
+      <div v-else class="files-table-wrapper">
+        <!-- Table Header -->
+        <div class="table-header">
+          <div class="table-select">
+            <el-checkbox
+              :model-value="isAllSelected"
+              @change="toggleSelectAll"
+              class="select-all"
+            >
+              <span v-if="selectedCount > 0" class="selected-text">已选择 {{ selectedCount }} 项</span>
+              <span v-else>全选</span>
+            </el-checkbox>
+          </div>
+          <div class="table-actions" v-if="selectedCount > 0">
+            <el-button type="danger" size="small" plain @click="clearSelection" class="batch-clear-btn">
+              <el-icon><Close /></el-icon>
+              <span>清除选择</span>
+            </el-button>
+            <el-button type="danger" size="small" plain @click="batchDelete" class="batch-delete-btn">
+              <el-icon><Delete /></el-icon>
+              <span>批量删除</span>
+            </el-button>
+          </div>
+        </div>
+
+        <!-- Table Body -->
+        <div class="files-table">
+          <div
+            v-for="(file, index) in filteredFiles"
+            :key="file.id"
+            class="file-row"
+            :class="{
+              'is-selected': isSelected(file.id),
+              'is-processing': file.status === 'processing',
+              'row-animated': isInitialLoad
+            }"
+            :style="{ '--delay': index * 0.04 + 's' }"
+            @click="toggleSelect(file.id)"
+          >
+            <!-- Select Checkbox -->
+            <div class="col-select" @click.stop>
+              <el-checkbox
+                :model-value="isSelected(file.id)"
+                @change="toggleSelect(file.id)"
+              />
+            </div>
+
+            <!-- File Icon -->
+            <div class="col-icon">
+              <div class="file-type-icon" :style="{ '--type-color': getTypeColor(file.file_type) }">
+                <el-icon size="18">
+                  <component :is="getFileIcon(file.file_type)" />
+                </el-icon>
+              </div>
+            </div>
+
+            <!-- File Name -->
+            <div class="col-name">
+              <span class="filename-text">{{ file.filename }}</span>
+              <span class="file-ext">{{ getFileExt(file.filename) }}</span>
+            </div>
+
+            <!-- Size -->
+            <div class="col-size">
+              {{ formatSize(file.size) }}
+            </div>
+
+            <!-- Date -->
+            <div class="col-date">
+              {{ formatDate(file.created_at) }}
+            </div>
+
+            <!-- Status -->
+            <div class="col-status">
+              <div class="status-pill" :class="'status-' + file.status">
+                <span class="status-dot"></span>
+                <span class="status-text">{{ getStatusText(file.status) }}</span>
+              </div>
+            </div>
+
+            <!-- Actions -->
+            <div class="col-actions" @click.stop>
+              <el-tooltip content="预览" placement="top" v-if="file.status === 'completed'">
+                <el-button text size="small" class="action-btn preview" @click="handlePreview(file)">
+                  <el-icon><View /></el-icon>
+                </el-button>
+              </el-tooltip>
+              <el-button text size="small" class="action-btn delete" @click="openDeleteDialog(file)">
+                <el-icon><Delete /></el-icon>
+              </el-button>
+            </div>
+          </div>
+        </div>
+      </div>
+    </div>
+
+    <!-- Upload Dialog -->
+    <el-dialog v-model="uploadDialogVisible" title="上传文件" width="520px" class="upload-dialog" :close-on-click-modal="false">
+      <div class="upload-area" @click="triggerUpload">
+        <el-upload
+          ref="uploadRef"
+          class="upload-component"
+          :auto-upload="false"
+          :limit="10"
+          :on-change="handleChange"
+          :on-remove="handleRemove"
+          :file-list="fileList"
+          drag
+          multiple
+          accept=".pdf,.docx,.doc,.xlsx,.xls,.csv,.epub,.md,.txt"
+          style="display: none;"
+        />
+        <div class="upload-content">
+          <div class="upload-illustration">
+            <div class="upload-ring"></div>
+            <div class="upload-core">
+              <el-icon size="32"><UploadFilled /></el-icon>
+            </div>
+          </div>
+          <div class="upload-text">
+            拖拽文件到此处，或 <em>点击选择</em>
+          </div>
+          <div class="upload-hint">
+            支持 PDF、DOCX、Excel、EPUB、Markdown 等格式
+          </div>
+        </div>
+      </div>
+
+      <!-- Selected Files -->
+      <div v-if="fileList.length > 0" class="selected-area">
+        <div class="selected-header">
+          <span>已选择 <strong>{{ fileList.length }}</strong> 个文件</span>
+          <el-button text size="small" @click="fileList = []">清空</el-button>
+        </div>
+        <div class="selected-list">
+          <div v-for="item in fileList" :key="item.uid" class="selected-item">
+            <el-icon size="14"><Document /></el-icon>
+            <span>{{ item.name }}</span>
+          </div>
+        </div>
+      </div>
+
+      <template #footer>
+        <el-button @click="uploadDialogVisible = false">取消</el-button>
+        <el-button type="primary" @click="submitUpload" :loading="uploading" :disabled="fileList.length === 0">
+          开始上传
+        </el-button>
+      </template>
+    </el-dialog>
+  </div>
+
+  <Teleport to="body">
+    <Transition name="fade">
+      <div
+        v-if="previewVisible"
+        class="preview-backdrop"
+        @click="previewVisible = false"
+      ></div>
+    </Transition>
+
+    <Transition name="slide-right">
+      <div v-if="previewVisible" class="preview-modal">
+        <div class="preview-header">
+          <div class="header-title">
+            <el-icon class="title-icon"><Document /></el-icon>
+            <span class="filename">{{ previewFile?.filename }}</span>
+          </div>
+          <el-button class="close-btn" text @click="previewVisible = false">
+            <el-icon><Close /></el-icon>
+          </el-button>
+        </div>
+
+        <div class="preview-tabs-wrapper">
+          <div class="preview-tabs">
+            <button
+              class="tab-item"
+              :class="{ active: previewMode === 'source' }"
+              @click="switchPreviewMode('source')"
+            >
+              源文件
+            </button>
+            <button
+              class="tab-item"
+              :class="{ active: previewMode === 'markdown' }"
+              @click="switchPreviewMode('markdown')"
+            >
+              Markdown
+            </button>
+            <div class="tab-indicator" :class="{ 'at-right': previewMode === 'markdown' }"></div>
+          </div>
+        </div>
+
+        <div class="preview-content" v-loading="previewLoading">
+          <iframe v-if="isPdfPreview && pdfDataUrl" :src="pdfDataUrl" class="pdf-viewer"></iframe>
+          <pre v-else-if="previewContent" class="code-content">{{ previewContent }}</pre>
+          <div v-else-if="!previewLoading && !isPdfPreview" class="preview-empty">
+            <el-icon size="32"><Document /></el-icon>
+            <span>暂无内容</span>
+          </div>
+        </div>
+      </div>
+    </Transition>
+  </Teleport>
+
+  <DeleteDialog
+    v-model:visible="deleteDialogVisible"
+    title="删除文件"
+    :item-name="pendingDeleteFile?.filename || ''"
+    detail-text="该操作会移除原始文件以及关联的处理结果，请确认当前项目内不再需要它。"
+    warning-text="删除后不可恢复，文件相关的分割结果和后续数据将一并失效。"
+    confirm-text="确认删除文件"
+    :loading="deletingFile"
+    @confirm="confirmDeleteFile"
+  />
+</template>
+
+<script lang="ts" src="../page-logic/ProjectFilePage.ts"></script>
+<style scoped src="../styles/pages/project-file.css"></style>

frontend/src/pages/ProjectPage.vue

@@ -32,13 +32,6 @@
           <span class="nav-dot"></span>
         </router-link>
       </nav>
-
-      <div class="sidebar-footer">
-        <router-link to="/" class="home-link">
-          <el-icon><HomeFilled /></el-icon>
-          <span>返回首页</span>
-        </router-link>
-      </div>
     </aside>
 
     <!-- Main Content -->
@@ -51,7 +44,7 @@
 <script setup>
 import { ref, computed, onMounted } from 'vue'
 import { useRoute, useRouter } from 'vue-router'
-import { projectApi } from '@/api'
+import { projectApi } from '@/core/api'
 import { ElMessage } from 'element-plus'
 
 const route = useRoute()
@@ -64,7 +57,7 @@ const project = ref({ name: '加载中...', description: '' })
 
 const navItems = [
   { path: 'files', label: '文件管理', icon: 'Folder' },
-  { path: 'split', label: '文本分割', icon: 'Operation' },
+  { path: 'split', label: '分割生成', icon: 'Operation' },
   { path: 'questions', label: '问答管理', icon: 'ChatDotSquare' },
   { path: 'datasets', label: '数据集', icon: 'Collection' },
   { path: 'eval', label: '评估系统', icon: 'DataAnalysis' },
@@ -76,7 +69,8 @@ const isActive = (path) => route.path.includes(path)
 const fetchProject = async () => {
   try {
     const res = await projectApi.get(id.value)
-    project.value = res.data
+    // New format: returns project directly
+    project.value = res
   } catch (error) {
     ElMessage.error('加载项目失败')
   }
@@ -295,37 +289,6 @@ onMounted(() => fetchProject())
 }
 
 /* Sidebar Footer */
-.sidebar-footer {
-  padding: 16px 20px;
-  border-top: 1px solid var(--border-subtle);
-}
-
-.home-link {
-  display: flex;
-  align-items: center;
-  gap: 12px;
-  padding: 12px 14px;
-  border-radius: var(--radius-md);
-  color: var(--text-tertiary);
-  font-size: 14px;
-  text-decoration: none;
-  transition: all var(--transition-fast);
-}
-
-.home-link:hover {
-  background: var(--bg-hover);
-  color: var(--text-primary);
-}
-
-.home-link span {
-  transition: opacity 0.2s ease;
-}
-
-.sidebar.collapsed .home-link span {
-  opacity: 0;
-  width: 0;
-  overflow: hidden;
-}
 
 /* Main Content */
 .main-content {
@@ -347,8 +310,7 @@ onMounted(() => fetchProject())
   }
 
   .sidebar .project-details,
-  .sidebar .nav-label,
-  .sidebar .home-link span {
+  .sidebar .nav-label {
     display: none;
   }
 

frontend/src/pages/ProjectQuestionPage.vue

@@ -0,0 +1,244 @@
+<template>
+  <div class="question-manage">
+    <!-- Header -->
+    <div class="page-header">
+      <div class="header-left">
+        <h2 class="page-title">问答管理</h2>
+        <p class="page-subtitle">管理和生成问答数据</p>
+      </div>
+      <div class="header-actions">
+        <el-button type="primary" @click="showGenerateDialog = true" class="generate-btn">
+          <el-icon><Plus /></el-icon>
+          <span>生成问题</span>
+        </el-button>
+      </div>
+    </div>
+
+    <!-- Stats Cards -->
+    <div class="stats-grid">
+      <div
+        class="stat-card stat-total"
+        :class="{ active: filterStatus === '' }"
+        @click="filterStatus = ''"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><ChatDotSquare /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ questions.length }}</span>
+            <span class="stat-label">总问题数</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-completed"
+        :class="{ active: filterStatus === 'generated' }"
+        @click="filterStatus = filterStatus === 'generated' ? '' : 'generated'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><MagicStick /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ generatedCount }}</span>
+            <span class="stat-label">AI 生成</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-processing"
+        :class="{ active: filterStatus === 'manual' }"
+        @click="filterStatus = filterStatus === 'manual' ? '' : 'manual'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><EditPen /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ manualCount }}</span>
+            <span class="stat-label">手动添加</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-failed"
+        :class="{ active: filterStatus === 'failed' }"
+        @click="filterStatus = filterStatus === 'failed' ? '' : 'failed'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><CircleCloseFilled /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ failedCount }}</span>
+            <span class="stat-label">失败</span>
+          </div>
+        </div>
+      </div>
+    </div>
+
+    <!-- Question Container -->
+    <div class="question-container" v-loading="loading && isInitialLoad">
+      <!-- Empty State -->
+      <div v-if="!loading && !isInitialLoad && filteredQuestions.length === 0" class="empty-state">
+        <div class="empty-illustration">
+          <div class="orbit orbit-1"></div>
+          <div class="orbit orbit-2"></div>
+          <div class="orbit orbit-3"></div>
+          <div class="empty-core">
+            <el-icon size="40"><ChatDotSquare /></el-icon>
+          </div>
+        </div>
+        <h3 class="empty-title">暂无问答数据</h3>
+        <p class="empty-desc">生成您的第一个问答数据集</p>
+        <el-button type="primary" @click="showGenerateDialog = true" class="empty-btn">生成问题</el-button>
+      </div>
+
+      <!-- Question Table -->
+      <div v-else class="question-table-wrapper">
+        <!-- Table Header -->
+        <div class="table-header">
+          <div class="table-select">
+            <el-checkbox
+              :model-value="isAllSelected"
+              @change="toggleSelectAll"
+              class="select-all"
+            >
+              <span v-if="selectedCount > 0" class="selected-text">已选择 {{ selectedCount }} 项</span>
+              <span v-else>全选</span>
+            </el-checkbox>
+          </div>
+          <div class="table-actions" v-if="selectedCount > 0">
+            <el-button type="danger" size="small" plain @click="clearSelection" class="batch-clear-btn">
+              <el-icon><Close /></el-icon>
+              <span>清除选择</span>
+            </el-button>
+            <el-button type="danger" size="small" plain @click="batchDelete" class="batch-delete-btn">
+              <el-icon><Delete /></el-icon>
+              <span>批量删除</span>
+            </el-button>
+          </div>
+        </div>
+
+        <!-- Table Body -->
+        <div class="question-table">
+          <div
+            v-for="(question, index) in filteredQuestions"
+            :key="question.id"
+            class="question-row"
+            :class="{
+              'is-selected': isSelected(question.id),
+              'row-animated': isInitialLoad
+            }"
+            :style="{ '--delay': index * 0.04 + 's' }"
+            @click="toggleSelect(question.id)"
+          >
+            <!-- Select Checkbox -->
+            <div class="col-select" @click.stop>
+              <el-checkbox
+                :model-value="isSelected(question.id)"
+                @change="toggleSelect(question.id)"
+              />
+            </div>
+
+            <!-- Question Content -->
+            <div class="col-content">
+              <div class="question-text">{{ question.content }}</div>
+              <div class="answer-text" v-if="question.answer">答: {{ question.answer }}</div>
+            </div>
+
+            <!-- Type -->
+            <div class="col-type">
+              <el-tag size="small" :style="{ '--tag-color': getTypeColor(question.question_type) }" effect="dark">
+                {{ getTypeName(question.question_type) }}
+              </el-tag>
+            </div>
+
+            <!-- Source -->
+            <div class="col-source">
+              <span class="source-badge" :class="'source-' + question.source">{{ getSourceName(question.source) }}</span>
+            </div>
+
+            <!-- Actions -->
+            <div class="col-actions" @click.stop>
+              <el-popconfirm title="确定删除此问题？" @confirm="handleDelete(question)">
+                <template #reference>
+                  <el-button text size="small" class="action-btn delete">
+                    <el-icon><Delete /></el-icon>
+                  </el-button>
+                </template>
+              </el-popconfirm>
+            </div>
+          </div>
+        </div>
+      </div>
+    </div>
+
+    <el-dialog v-model="showGenerateDialog" title="生成问题" width="640px" class="generate-dialog">
+      <el-form :model="generateConfig" label-position="top">
+        <el-form-item label="生成模型">
+          <el-select
+            v-model="generateConfig.model_id"
+            placeholder="选择 chat / vlm 模型"
+            style="width: 100%"
+            size="large"
+          >
+            <el-option
+              v-for="model in generateModels"
+              :key="model.id"
+              :label="`${model.model_name} · ${getProviderLabel(model.provider)}`"
+              :value="model.id"
+            />
+          </el-select>
+        </el-form-item>
+        <el-form-item label="文本块">
+          <el-select
+            v-model="generateConfig.chunk_ids"
+            multiple
+            placeholder="选择文本块"
+            style="width: 100%"
+            size="large"
+          >
+            <el-option
+              v-for="chunk in chunks"
+              :key="chunk.id"
+              :label="chunk.name || chunk.content.slice(0, 50) + '...'"
+              :value="chunk.id"
+            />
+          </el-select>
+        </el-form-item>
+        <div class="form-row">
+          <el-form-item label="每个块生成数量">
+            <el-input-number v-model="generateConfig.count" :min="1" :max="8" size="large" style="width: 100%" />
+          </el-form-item>
+        </div>
+        <el-form-item label="生成策略">
+          <div style="display: flex; gap: 16px; flex-wrap: wrap;">
+            <el-checkbox v-model="generateConfig.dirty_data_filter">脏数据过滤</el-checkbox>
+            <el-checkbox v-model="generateConfig.thinking_mode">思考模式</el-checkbox>
+          </div>
+        </el-form-item>
+        <el-form-item label="预设提示语">
+          <el-input
+            v-model="generateConfig.preset_prompt"
+            type="textarea"
+            :rows="8"
+            resize="none"
+          />
+        </el-form-item>
+      </el-form>
+      <template #footer>
+        <el-button @click="showGenerateDialog = false">取消</el-button>
+        <el-button type="primary" @click="handleGenerate" :loading="generating">开始生成</el-button>
+      </template>
+    </el-dialog>
+  </div>
+</template>
+
+<script lang="ts" src="../page-logic/ProjectQuestionPage.ts"></script>
+<style scoped src="../styles/pages/project-question.css"></style>

frontend/src/pages/ProjectSettingsPage.vue

@@ -95,7 +95,7 @@
 import { ref, reactive, onMounted, computed } from 'vue'
 import { useRoute } from 'vue-router'
 import { ElMessage } from 'element-plus'
-import { projectApi } from '@/api'
+import { projectApi } from '@/core/api'
 
 const route = useRoute()
 const projectId = computed(() => route.params.id)
@@ -125,8 +125,9 @@ const prompts = reactive({
 const fetchProject = async () => {
   try {
     const res = await projectApi.get(projectId.value)
-    projectInfo.name = res.data.name
-    projectInfo.description = res.data.description || ''
+    // New format: project directly in response
+    projectInfo.name = res.name
+    projectInfo.description = res.description || ''
   } catch (error) {
     console.error(error)
   }

frontend/src/pages/ProjectTextSplitPage.vue

@@ -0,0 +1,830 @@
+<template>
+  <div class="text-split">
+    <!-- Header -->
+    <div class="page-header">
+      <div class="header-left">
+        <h2 class="page-title">分割生成</h2>
+        <p class="page-subtitle">选择文件进行智能分割</p>
+      </div>
+      <div class="header-actions">
+        <el-button
+          @click="openSplitDialog"
+          :disabled="selectedFiles.length === 0"
+          class="split-btn"
+        >
+          <el-icon><ChatDotSquare /></el-icon>
+          <span>批量分割</span>
+        </el-button>
+        <el-button
+          type="primary"
+          @click="openGenerateDialog"
+          :disabled="completedFiles === 0"
+          class="generate-btn"
+        >
+          <el-icon><VideoPlay /></el-icon>
+          <span>批量生成</span>
+        </el-button>
+      </div>
+    </div>
+
+    <!-- Stats Cards -->
+    <div class="stats-grid">
+      <div
+        class="stat-card stat-total"
+        :class="{ active: filterStatus === '' }"
+        @click="filterStatus = ''"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><Document /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ files.length }}</span>
+            <span class="stat-label">总文件</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-completed"
+        :class="{ active: filterStatus === 'completed' }"
+        @click="filterStatus = filterStatus === 'completed' ? '' : 'completed'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><CircleCheckFilled /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ completedFiles }}</span>
+            <span class="stat-label">已分割</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-processing"
+        :class="{ active: filterStatus === 'processing' }"
+        @click="filterStatus = filterStatus === 'processing' ? '' : 'processing'"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><Loading /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ processingCount }}</span>
+            <span class="stat-label">分割中</span>
+          </div>
+        </div>
+      </div>
+      <div
+        class="stat-card stat-chunks"
+      >
+        <div class="stat-glow"></div>
+        <div class="stat-inner">
+          <div class="stat-icon-wrap">
+            <el-icon size="24"><List /></el-icon>
+          </div>
+          <div class="stat-info">
+            <span class="stat-value">{{ totalChunks }}</span>
+            <span class="stat-label">总文本块</span>
+          </div>
+        </div>
+      </div>
+    </div>
+
+    <!-- File List / Split View -->
+    <div class="content-area">
+      <!-- Empty State -->
+      <div v-if="!loading && !isInitialLoad && filteredFiles.length === 0" class="empty-state">
+        <div class="empty-illustration">
+          <div class="orbit orbit-1"></div>
+          <div class="orbit orbit-2"></div>
+          <div class="orbit orbit-3"></div>
+          <div class="empty-core">
+            <el-icon size="40"><FolderOpened /></el-icon>
+          </div>
+        </div>
+        <h3 class="empty-title">暂无可分割文件</h3>
+        <p class="empty-desc">请先在文件管理中上传文档</p>
+      </div>
+
+      <!-- File Table -->
+      <div v-else class="files-table-wrapper">
+        <div class="table-header">
+          <div class="table-select">
+            <el-checkbox
+              :model-value="isAllSelected"
+              @change="toggleSelectAll"
+              class="select-all"
+            >
+              <span v-if="selectedCount > 0" class="selected-text">已选择 {{ selectedCount }} 项</span>
+              <span v-else>全选</span>
+            </el-checkbox>
+          </div>
+          <div class="table-actions" v-if="selectedCount > 0">
+            <el-button type="danger" size="small" plain @click="clearSelection" class="batch-clear-btn">
+              <el-icon><Close /></el-icon>
+              <span>清除选择</span>
+            </el-button>
+          </div>
+        </div>
+        <div class="files-list">
+          <div
+            v-for="(file, index) in filteredFiles"
+            :key="file.id"
+            class="file-row"
+            :class="{
+              'is-selected': isSelected(file.id),
+              'is-processing': file.status === 'processing'
+            }"
+            :style="{ '--delay': index * 0.04 + 's' }"
+            @click="toggleSelect(file.id)"
+          >
+            <div class="col-select">
+              <el-checkbox
+                :model-value="isSelected(file.id)"
+                @click.stop
+                @change="toggleSelect(file.id)"
+              />
+            </div>
+            <div class="col-icon">
+              <div class="file-type-icon" style="background: #8b5cf6;">
+                <el-icon size="18" color="white">
+                  <Document />
+                </el-icon>
+              </div>
+            </div>
+            <div class="col-name">
+              <span class="file-name">{{ file.filename }}.md</span>
+              <span class="file-meta">{{ formatSize(file.size) }}</span>
+            </div>
+            <div class="col-chunks">
+              <span v-if="fileChunks[file.id]" class="chunk-count">
+                {{ fileChunks[file.id] }} 块
+              </span>
+              <span v-else class="chunk-count empty">-</span>
+            </div>
+            <div class="col-status">
+              <div v-if="file.status === 'processing'" class="status-badge processing">
+                <el-icon class="spin" size="12"><Loading /></el-icon>
+                <span>分割中</span>
+              </div>
+              <div v-else-if="fileChunks[file.id]" class="status-badge success">
+                <el-icon size="12"><CircleCheckFilled /></el-icon>
+                <span>已完成</span>
+              </div>
+              <div v-else class="status-badge pending">
+                <el-icon size="12"><Clock /></el-icon>
+                <span>待分割</span>
+              </div>
+            </div>
+            <div class="col-operations" v-if="fileChunks[file.id]">
+              <el-tooltip content="预览修改" placement="top">
+                <el-button text size="small" class="op-btn" @click.stop="openChunkPreview(file)">
+                  <el-icon><Edit /></el-icon>
+                </el-button>
+              </el-tooltip>
+              <el-tooltip content="删除该文件的所有块" placement="top">
+                <el-button
+                  text
+                  size="small"
+                  class="op-btn delete"
+                  :loading="deletingFileChunksId === file.id"
+                  @click.stop="openDeleteFileChunksDialog(file)"
+                >
+                  <el-icon><Delete /></el-icon>
+                </el-button>
+              </el-tooltip>
+            </div>
+          </div>
+        </div>
+      </div>
+    </div>
+  </div>
+
+  <!-- Batch Split Dialog -->
+  <Teleport to="body">
+    <Transition name="dialog-fade">
+      <div v-if="splitDialogVisible" class="split-dialog-overlay" @click.self="splitDialogVisible = false">
+        <div class="split-dialog">
+          <!-- Header -->
+          <div class="dialog-header">
+            <div class="header-icon">
+              <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                <path d="M12 2L2 7l10 5 10-5-10-5z"/>
+                <path d="M2 17l10 5 10-5"/>
+                <path d="M2 12l10 5 10-5"/>
+              </svg>
+            </div>
+            <div class="header-text">
+              <h3>批量分割配置</h3>
+              <span class="header-sub">{{ selectedFiles.length }} 个文件待处理</span>
+            </div>
+            <button class="close-btn" @click="splitDialogVisible = false">
+              <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                <path d="M18 6L6 18M6 6l12 12"/>
+              </svg>
+            </button>
+          </div>
+
+          <!-- Dialog Body -->
+          <div class="dialog-body">
+            <!-- Split Method Selector -->
+            <div class="method-selector">
+              <div class="method-label">
+                <span class="label-icon">◆</span>
+                <span>分割算法</span>
+              </div>
+              <div class="method-grid">
+                <button
+                  v-for="m in methods"
+                  :key="m.value"
+                  class="method-btn"
+                  :class="{ active: splitConfig.method === m.value }"
+                  @click="splitConfig.method = m.value"
+                >
+                  <span class="method-name">{{ m.label }}</span>
+                  <span class="method-tag">{{ m.tag }}</span>
+                </button>
+              </div>
+            </div>
+
+            <!-- Parameters Panel -->
+            <div class="params-panel">
+              <!-- Panel Grid Background -->
+              <div class="panel-grid"></div>
+
+              <!-- Common Parameters -->
+              <div class="param-section" v-if="splitConfig.method !== 'paragraph'">
+                <div class="section-header">
+                  <span class="section-num">01</span>
+                  <span class="section-title">块大小控制</span>
+                </div>
+                <div class="param-row">
+                  <div class="param-info">
+                    <span class="param-name">chunk_size</span>
+                    <span class="param-desc">每个文本块的字符数</span>
+                  </div>
+                  <div class="param-control">
+                    <input
+                      type="range"
+                      v-model.number="splitConfig.chunk_size"
+                      :min="100"
+                      :max="2000"
+                      :step="100"
+                      class="cyber-slider"
+                    />
+                    <div class="param-value">
+                      <span class="value-num">{{ splitConfig.chunk_size }}</span>
+                      <span class="value-unit">chars</span>
+                    </div>
+                  </div>
+                </div>
+                <!-- Visual Preview -->
+                <div class="chunk-preview">
+                  <div class="preview-label">预览</div>
+                  <div class="preview-bars">
+                    <div
+                      v-for="i in 5"
+                      :key="i"
+                      class="preview-bar"
+                      :style="{ width: (splitConfig.chunk_size / 20) + 'px' }"
+                    ></div>
+                  </div>
+                </div>
+              </div>
+
+              <div class="param-section" v-if="splitConfig.method !== 'sentence'">
+                <div class="section-header">
+                  <span class="section-num">02</span>
+                  <span class="section-title">重叠控制</span>
+                </div>
+                <div class="param-row">
+                  <div class="param-info">
+                    <span class="param-name">overlap</span>
+                    <span class="param-desc">相邻块之间的重叠字符数</span>
+                  </div>
+                  <div class="param-control">
+                    <input
+                      type="range"
+                      v-model.number="splitConfig.overlap"
+                      :min="0"
+                      :max="500"
+                      :step="50"
+                      class="cyber-slider"
+                    />
+                    <div class="param-value">
+                      <span class="value-num">{{ splitConfig.overlap }}</span>
+                      <span class="value-unit">chars</span>
+                    </div>
+                  </div>
+                </div>
+                <!-- Overlap Visual -->
+                <div class="overlap-preview">
+                  <div class="overlap-block" :style="{ width: (splitConfig.chunk_size / 4) + 'px' }"></div>
+                  <div class="overlap-zone" :style="{ width: (splitConfig.overlap / 4) + 'px' }"></div>
+                  <div class="overlap-block" :style="{ width: (splitConfig.chunk_size / 4) + 'px' }"></div>
+                </div>
+              </div>
+
+              <!-- Custom Separator -->
+              <div class="param-section" v-if="splitConfig.method === 'custom'">
+                <div class="section-header">
+                  <span class="section-num">03</span>
+                  <span class="section-title">自定义分隔符</span>
+                </div>
+                <div class="param-row">
+                  <div class="param-info">
+                    <span class="param-name">separator</span>
+                    <span class="param-desc">用于分割文本的字符序列</span>
+                  </div>
+                  <div class="param-control full">
+                    <input
+                      type="text"
+                      v-model="splitConfig.separator"
+                      placeholder="\n\n"
+                      class="cyber-input"
+                    />
+                  </div>
+                </div>
+              </div>
+
+              <template v-if="splitConfig.method === 'semantic'">
+                <div class="param-section">
+                  <div class="section-header highlight">
+                    <span class="section-num">04</span>
+                    <span class="section-title">句段优先说明</span>
+                    <span class="section-badge">RULE</span>
+                  </div>
+                  <div class="param-row full">
+                    <div class="param-info">
+                      <span class="param-name">规则型切分</span>
+                      <span class="param-desc">按段落和句子边界优先切分，不调用 embedding API，更接近“句段优先的递归切分”。</span>
+                    </div>
+                  </div>
+                </div>
+              </template>
+              <template v-if="splitConfig.method === 'semantic_embedding'">
+                <div class="param-section">
+                  <div class="section-header highlight">
+                    <span class="section-num">04</span>
+                    <span class="section-title">Embedding 模型</span>
+                    <span class="section-badge">API</span>
+                  </div>
+                  <div class="api-grid">
+                    <div class="param-row full">
+                      <div class="param-info">
+                        <span class="param-name">embedding_model</span>
+                        <span class="param-desc">直接使用模型管理中已配置的 embedding 模型</span>
+                      </div>
+                      <div class="param-control full">
+                        <select v-model="splitConfig.embedding_model_id" class="cyber-select" :disabled="embeddingModels.length === 0">
+                          <option value="" disabled>
+                            {{ embeddingModels.length ? '选择 embedding 模型' : '暂无可用 embedding 模型' }}
+                          </option>
+                          <option
+                            v-for="model in embeddingModels"
+                            :key="model.id"
+                            :value="model.id"
+                          >
+                            {{ model.model_name }} · {{ getProviderLabel(model.provider) }}
+                          </option>
+                        </select>
+                      </div>
+                    </div>
+                    <div class="param-row full" v-if="embeddingModels.length === 0">
+                      <div class="embedding-empty-hint">
+                        <span>还没有可用的 embedding 模型。</span>
+                        <button type="button" class="text-link-btn" @click="goToModelSettings">
+                          去模型配置
+                        </button>
+                      </div>
+                    </div>
+                  </div>
+                </div>
+
+                <div class="param-section">
+                  <div class="section-header highlight">
+                    <span class="section-num">05</span>
+                    <span class="section-title">语义边界参数</span>
+                    <span class="section-badge">AI</span>
+                  </div>
+                  <div class="param-row">
+                    <div class="param-info">
+                      <span class="param-name">similarity_threshold</span>
+                      <span class="param-desc">越高越保守，越低越容易切出新块</span>
+                    </div>
+                    <div class="param-control">
+                      <input
+                        type="range"
+                        v-model.number="splitConfig.similarity_threshold"
+                        :min="0.1"
+                        :max="0.9"
+                        :step="0.05"
+                        class="cyber-slider accent"
+                      />
+                      <div class="param-value accent">
+                        <span class="value-num">{{ splitConfig.similarity_threshold.toFixed(2) }}</span>
+                      </div>
+                    </div>
+                  </div>
+                  <div class="param-row">
+                    <div class="param-info">
+                      <span class="param-name">min_chunk_size</span>
+                      <span class="param-desc">最小块大小，避免语义过碎</span>
+                    </div>
+                    <div class="param-control">
+                      <input
+                        type="range"
+                        v-model.number="splitConfig.min_chunk_size"
+                        :min="50"
+                        :max="500"
+                        :step="10"
+                        class="cyber-slider accent"
+                      />
+                      <div class="param-value accent">
+                        <span class="value-num">{{ splitConfig.min_chunk_size }}</span>
+                        <span class="value-unit">chars</span>
+                      </div>
+                    </div>
+                  </div>
+                </div>
+              </template>
+
+            </div>
+          </div>
+
+          <div class="dialog-footer">
+            <button class="btn-cancel" @click="splitDialogVisible = false">
+              <span>取消</span>
+            </button>
+            <button class="btn-confirm" @click="handleBatchSplit" :class="{ loading: splitting }">
+              <span v-if="!splitting" class="btn-text">
+                <svg class="btn-icon" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                  <polygon points="5 3 19 12 5 21 5 3"/>
+                </svg>
+                开始生成
+              </span>
+              <span v-else class="btn-loading">
+                <span class="loading-dot"></span>
+                <span class="loading-dot"></span>
+                <span class="loading-dot"></span>
+              </span>
+            </button>
+          </div>
+        </div>
+      </div>
+    </Transition>
+  </Teleport>
+
+  <Teleport to="body">
+    <Transition name="dialog-fade">
+      <div v-if="generateDialogVisible" class="split-dialog-overlay" @click.self="generateDialogVisible = false">
+        <div class="split-dialog generate-dialog-shell">
+          <div class="dialog-header">
+            <div class="header-icon">
+              <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                <path d="M12 2l2.4 5.8L20 10l-4 4 1 6-5-3-5 3 1-6-4-4 5.6-2.2L12 2z"/>
+              </svg>
+            </div>
+            <div class="header-text">
+              <h3>批量生成问答</h3>
+              <span class="header-sub">面向当前项目全部已分割文本块</span>
+            </div>
+            <button class="close-btn" @click="generateDialogVisible = false">
+              <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                <path d="M18 6L6 18M6 6l12 12"/>
+              </svg>
+            </button>
+          </div>
+
+          <div class="dialog-body">
+            <div class="params-panel">
+              <div class="panel-grid"></div>
+
+              <div class="param-section">
+                <div class="section-header">
+                  <span class="section-num">01</span>
+                  <span class="section-title">大语言模型</span>
+                </div>
+                <div class="param-row full">
+                  <div class="param-info">
+                    <span class="param-name">model</span>
+                    <span class="param-desc">选择用于生成问答对的 chat / vlm 模型</span>
+                  </div>
+                  <div class="param-control full">
+                    <el-select
+                      v-model="generateConfig.model_id"
+                      class="cyber-select generate-model-select"
+                      :disabled="generateModels.length === 0"
+                      :teleported="false"
+                      placement="bottom-start"
+                      popper-class="generate-model-dropdown"
+                      placeholder="选择生成模型"
+                    >
+                      <el-option
+                        v-for="model in generateModels"
+                        :key="model.id"
+                        :label="`${model.model_name} · ${getProviderLabel(model.provider)}`"
+                        :value="model.id"
+                      >
+                        <div class="generate-option">
+                          <div class="generate-option__title">{{ model.model_name }}</div>
+                          <div class="generate-option__meta">{{ getProviderLabel(model.provider) }}</div>
+                        </div>
+                      </el-option>
+                    </el-select>
+                  </div>
+                </div>
+                <div class="param-row full" v-if="generateModels.length === 0">
+                  <div class="embedding-empty-hint">
+                    <span>还没有可用的 chat / vlm 模型。</span>
+                    <button type="button" class="text-link-btn" @click="goToModelSettings">
+                      去模型配置
+                    </button>
+                  </div>
+                </div>
+              </div>
+
+              <div class="param-section generate-strategy-section">
+                <div class="section-header">
+                  <span class="section-num">02</span>
+                  <span class="section-title">生成策略</span>
+                </div>
+                <div class="generate-strategy-grid">
+                  <button
+                    type="button"
+                    class="strategy-card"
+                    :class="{ active: generateConfig.dirty_data_filter }"
+                    @click="generateConfig.dirty_data_filter = !generateConfig.dirty_data_filter"
+                  >
+                    <div class="strategy-card__head">
+                      <span class="strategy-card__title">脏数据过滤</span>
+                      <span class="strategy-card__switch" :class="{ active: generateConfig.dirty_data_filter }">
+                        <span class="strategy-card__switch-handle"></span>
+                      </span>
+                    </div>
+                    <p class="strategy-card__desc">过滤目录、极短内容和明显噪声，减少无效调用。</p>
+                    <span class="strategy-card__state">{{ generateConfig.dirty_data_filter ? '已开启' : '已关闭' }}</span>
+                  </button>
+                  <button
+                    type="button"
+                    class="strategy-card"
+                    :class="{ active: generateConfig.thinking_mode }"
+                    @click="generateConfig.thinking_mode = !generateConfig.thinking_mode"
+                  >
+                    <div class="strategy-card__head">
+                      <span class="strategy-card__title">思考模式</span>
+                      <span class="strategy-card__switch" :class="{ active: generateConfig.thinking_mode }">
+                        <span class="strategy-card__switch-handle"></span>
+                      </span>
+                    </div>
+                    <p class="strategy-card__desc">生成前强化内容分析，提升问题质量与覆盖度。</p>
+                    <span class="strategy-card__state">{{ generateConfig.thinking_mode ? '已开启' : '已关闭' }}</span>
+                  </button>
+                </div>
+                <div class="param-row">
+                  <div class="param-info">
+                    <span class="param-name">single_chunk_count</span>
+                    <span class="param-desc">每个 chunk 生成多少组问答</span>
+                  </div>
+                  <div class="param-control">
+                    <input
+                      type="range"
+                      v-model.number="generateConfig.count"
+                      :min="1"
+                      :max="8"
+                      :step="1"
+                      class="cyber-slider accent"
+                    />
+                    <div class="param-value accent">
+                      <span class="value-num">{{ generateConfig.count }}</span>
+                      <span class="value-unit">pairs</span>
+                    </div>
+                  </div>
+                </div>
+              </div>
+
+              <div class="param-section generate-prompt-section">
+                <div class="section-header">
+                  <span class="section-num">03</span>
+                  <span class="section-title">预设提示语</span>
+                </div>
+                <div class="generate-prompt-box">
+                  <textarea
+                    v-model="generateConfig.preset_prompt"
+                    class="cyber-textarea generate-prompt-textarea"
+                    rows="9"
+                  />
+                </div>
+              </div>
+            </div>
+          </div>
+
+          <div class="dialog-footer">
+            <button class="btn-cancel" @click="generateDialogVisible = false">
+              <span>取消</span>
+            </button>
+            <button class="btn-confirm" @click="handleBatchGenerate" :class="{ loading: generatingQuestions }">
+              <span v-if="!generatingQuestions" class="btn-text">
+                <svg class="btn-icon" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                  <polygon points="5 3 19 12 5 21 5 3"/>
+                </svg>
+                开始生成
+              </span>
+              <span v-else class="btn-loading">
+                <span class="loading-dot"></span>
+                <span class="loading-dot"></span>
+                <span class="loading-dot"></span>
+              </span>
+            </button>
+          </div>
+        </div>
+      </div>
+    </Transition>
+  </Teleport>
+
+  <Teleport to="body">
+    <Transition name="dialog-fade">
+      <div v-if="chunkPreviewVisible" class="chunk-preview-overlay" @click.self="chunkPreviewVisible = false">
+        <div class="chunk-preview-dialog">
+          <div class="dialog-header">
+            <div class="header-icon">
+              <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                <path d="M12 2L2 7l10 5 10-5-10-5z"/>
+                <path d="M2 17l10 5 10-5"/>
+                <path d="M2 12l10 5 10-5"/>
+              </svg>
+            </div>
+            <div class="header-text">
+              <h3>预览修改</h3>
+              <span class="header-sub">{{ previewFile?.filename }} - {{ previewChunks.length }} 个块</span>
+            </div>
+            <button class="close-btn" @click="chunkPreviewVisible = false">
+              <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
+                <path d="M18 6L6 18M6 6l12 12"/>
+              </svg>
+            </button>
+          </div>
+
+          <div class="chunk-preview-toolbar">
+            <label class="preview-search">
+              <span class="preview-search-label">检索</span>
+              <input
+                v-model.trim="previewSearch"
+                type="text"
+                class="cyber-input"
+                placeholder="搜索块内容、块号"
+              />
+            </label>
+
+            <div class="preview-filter-group">
+              <button
+                class="preview-filter-btn"
+                :class="{ active: previewFilter === 'all' }"
+                @click="previewFilter = 'all'"
+              >
+                全部
+              </button>
+              <button
+                class="preview-filter-btn"
+                :class="{ active: previewFilter === 'modified' }"
+                @click="previewFilter = 'modified'"
+              >
+                仅已修改
+              </button>
+            </div>
+
+            <label class="preview-jump">
+              <span class="preview-search-label">跳转</span>
+              <input
+                v-model.trim="previewJumpInput"
+                type="text"
+                inputmode="numeric"
+                class="cyber-input"
+                placeholder="块号"
+                @keydown.enter.prevent="jumpToChunk"
+              />
+            </label>
+
+            <div class="preview-stats">
+              <span>{{ filteredPreviewChunks.length }} / {{ previewChunks.length }} 块</span>
+              <span>{{ modifiedPreviewCount }} 已修改</span>
+            </div>
+          </div>
+
+          <div class="dialog-body chunk-workspace">
+            <div v-if="previewLoading" class="loading-state">
+              <span class="loading-dot"></span>
+              <span class="loading-dot"></span>
+              <span class="loading-dot"></span>
+            </div>
+            <div v-else-if="previewChunks.length" class="chunk-workspace-grid">
+              <aside class="chunk-nav-panel">
+                <div class="chunk-nav-list">
+                  <button
+                    v-for="chunk in filteredPreviewChunks"
+                    :key="chunk.id"
+                    class="chunk-nav-item"
+                    :class="{
+                      active: chunk.id === selectedPreviewChunkId,
+                      modified: isChunkModified(chunk)
+                    }"
+                    @click="selectPreviewChunk(chunk.id)"
+                  >
+                    <div class="chunk-nav-meta">
+                      <span class="chunk-nav-index">块 {{ chunk.displayIndex }}</span>
+                      <span class="chunk-nav-words">{{ chunk.word_count || 0 }} 字</span>
+                    </div>
+                    <p class="chunk-nav-snippet">{{ getChunkSnippet(chunk) }}</p>
+                    <div class="chunk-nav-footer">
+                      <span class="chunk-nav-status">{{ isChunkModified(chunk) ? '已修改' : '原始' }}</span>
+                    </div>
+                  </button>
+                </div>
+              </aside>
+
+              <section v-if="activePreviewChunk" class="chunk-editor-panel">
+                <div class="chunk-item">
+                  <div class="chunk-header">
+                    <div class="chunk-header-main">
+                      <span class="chunk-index">块 {{ activePreviewChunkIndex }} / {{ previewChunks.length }}</span>
+                      <span class="chunk-state-pill" :class="{ modified: isChunkModified(activePreviewChunk) }">
+                        {{ isChunkModified(activePreviewChunk) ? '已修改未保存' : '内容未变更' }}
+                      </span>
+                    </div>
+                    <span class="chunk-words">{{ activePreviewChunk.word_count || 0 }} 字</span>
+                  </div>
+                  <div class="chunk-form">
+                    <div class="form-group">
+                      <label>内容</label>
+                      <textarea
+                        v-model="activePreviewChunk.editingContent"
+                        class="cyber-textarea preview-editor"
+                        rows="20"
+                      ></textarea>
+                    </div>
+                    <div class="chunk-actions">
+                      <button
+                        class="btn-secondary"
+                        :disabled="!isChunkModified(activePreviewChunk)"
+                        @click="resetChunk(activePreviewChunk)"
+                      >
+                        还原
+                      </button>
+                      <button
+                        class="btn-save"
+                        :class="{ loading: savingChunks }"
+                        :disabled="!isChunkModified(activePreviewChunk)"
+                        @click="saveChunk(activePreviewChunk)"
+                      >
+                        <span v-if="!savingChunks">保存</span>
+                        <span v-else class="btn-loading">
+                          <span class="loading-dot"></span>
+                          <span class="loading-dot"></span>
+                          <span class="loading-dot"></span>
+                        </span>
+                      </button>
+                      <button
+                        class="btn-delete"
+                        :class="{ loading: deletingChunkId === activePreviewChunk.id }"
+                        :disabled="deletingChunkId === activePreviewChunk.id"
+                        @click="openDeleteChunkDialog(activePreviewChunk)"
+                      >
+                        <span v-if="deletingChunkId !== activePreviewChunk.id">删除</span>
+                        <span v-else class="btn-loading">
+                          <span class="loading-dot"></span>
+                          <span class="loading-dot"></span>
+                          <span class="loading-dot"></span>
+                        </span>
+                      </button>
+                    </div>
+                  </div>
+                </div>
+              </section>
+            </div>
+            <div v-else class="chunk-empty-state">
+              <h4>没有匹配的分片</h4>
+              <p>试试清空搜索词，或切回“全部”查看所有分片。</p>
+            </div>
+          </div>
+        </div>
+      </div>
+    </Transition>
+  </Teleport>
+
+  <DeleteDialog
+    v-model:visible="deleteDialogVisible"
+    :title="deleteDialogTitle"
+    :item-name="deleteDialogItemName"
+    :detail-text="deleteDialogDetail"
+    :warning-text="deleteDialogWarning"
+    :confirm-text="deleteDialogConfirmText"
+    :loading="deleteDialogLoading"
+    @confirm="confirmDeleteAction"
+  />
+</template>
+
+<script lang="ts" src="../page-logic/ProjectTextSplitPage.ts"></script>
+<style scoped src="../styles/pages/project-text-split.css"></style>

frontend/src/router/index.js

@@ -1,67 +0,0 @@
-import { createRouter, createWebHistory } from 'vue-router'
-
-const routes = [
-  {
-    path: '/',
-    name: 'Home',
-    component: () => import('@/views/HomeView.vue')
-  },
-  {
-    path: '/project/:id',
-    name: 'Project',
-    component: () => import('@/views/ProjectView.vue'),
-    children: [
-      {
-        path: '',
-        redirect: to => `/project/${to.params.id}/files`
-      },
-      {
-        path: 'files',
-        name: 'ProjectFiles',
-        component: () => import('@/views/project/FileManage.vue')
-      },
-      {
-        path: 'split',
-        name: 'ProjectSplit',
-        component: () => import('@/views/project/TextSplit.vue')
-      },
-      {
-        path: 'questions',
-        name: 'ProjectQuestions',
-        component: () => import('@/views/project/QuestionManage.vue')
-      },
-      {
-        path: 'datasets',
-        name: 'ProjectDatasets',
-        component: () => import('@/views/project/DatasetManage.vue')
-      },
-      {
-        path: 'eval',
-        name: 'ProjectEval',
-        component: () => import('@/views/project/EvalManage.vue')
-      },
-      {
-        path: 'settings',
-        name: 'ProjectSettings',
-        component: () => import('@/views/project/Settings.vue')
-      }
-    ]
-  },
-  {
-    path: '/playground',
-    name: 'Playground',
-    component: () => import('@/views/PlaygroundView.vue')
-  },
-  {
-    path: '/data-square',
-    name: 'DataSquare',
-    component: () => import('@/views/DataSquareView.vue')
-  }
-]
-
-const router = createRouter({
-  history: createWebHistory(),
-  routes
-})
-
-export default router

frontend/src/shared/components/common/CreateProjectDialog.vue

@@ -0,0 +1,445 @@
+<template>
+  <el-dialog
+    :model-value="visible"
+    title=""
+    width="480px"
+    class="create-dialog"
+    :show-close="false"
+    align-center
+    @update:model-value="$emit('update:visible', $event)"
+  >
+    <template #header>
+      <div class="dialog-header">
+        <div class="header-glow"></div>
+        <div class="header-content">
+          <div class="dialog-icon-new">
+            <el-icon size="24"><FolderAdd /></el-icon>
+          </div>
+          <div class="header-text">
+            <h3>创建新项目</h3>
+            <p>开始构建您的AI训练数据集</p>
+          </div>
+        </div>
+        <button class="close-btn" @click="handleClose">
+          <el-icon><Close /></el-icon>
+        </button>
+      </div>
+    </template>
+    <div class="dialog-body">
+      <div class="form-card">
+        <div class="input-group">
+          <label class="input-label">
+            <span class="label-icon">✦</span>
+            项目名称
+          </label>
+          <el-input
+            v-model="formData.name"
+            placeholder="例如：客服问答数据集"
+            size="large"
+            class="custom-input"
+          />
+        </div>
+        <div class="input-group">
+          <label class="input-label">
+            <span class="label-icon">✦</span>
+            项目描述
+          </label>
+          <el-input
+            v-model="formData.description"
+            type="textarea"
+            :rows="3"
+            placeholder="描述这个项目的用途和数据类型..."
+            class="custom-input"
+          />
+        </div>
+      </div>
+
+      <!-- Quick Templates -->
+      <div class="templates-section">
+        <span class="templates-label">快速开始模板</span>
+        <div class="templates-grid">
+          <div
+            class="template-card"
+            :class="{ active: formData.type === 'qa' }"
+            @click="useTemplate('qa')"
+          >
+            <el-icon><ChatDotRound /></el-icon>
+            <span>问答对</span>
+          </div>
+          <div
+            class="template-card"
+            :class="{ active: formData.type === 'table' }"
+            @click="useTemplate('table')"
+          >
+            <el-icon><Document /></el-icon>
+            <span>表格</span>
+          </div>
+          <div
+            class="template-card"
+            :class="{ active: formData.type === 'database' }"
+            @click="useTemplate('database')"
+          >
+            <el-icon><Connection /></el-icon>
+            <span>数据库</span>
+          </div>
+        </div>
+      </div>
+    </div>
+    <template #footer>
+      <div class="dialog-footer">
+        <el-button
+          type="primary"
+          :loading="loading"
+          @click="handleSubmit"
+          class="btn-create"
+        >
+          <el-icon v-if="!loading"><Plus /></el-icon>
+          创建项目
+        </el-button>
+      </div>
+    </template>
+  </el-dialog>
+</template>
+
+<script setup>
+import { reactive, watch } from 'vue'
+
+const props = defineProps({
+  visible: {
+    type: Boolean,
+    default: false
+  },
+  loading: {
+    type: Boolean,
+    default: false
+  }
+})
+
+const emit = defineEmits(['update:visible', 'submit'])
+
+const formData = reactive({
+  name: '',
+  description: '',
+  type: ''
+})
+
+const templates = {
+  qa: { name: '问答数据集', description: '基于文档生成问答对训练数据' },
+  table: { name: '表格数据集', description: '从表格数据生成结构化训练数据' },
+  database: { name: '数据库数据集', description: '从数据库导出数据生成训练数据' }
+}
+
+const useTemplate = (type) => {
+  formData.type = type
+}
+
+const handleClose = () => {
+  emit('update:visible', false)
+}
+
+const handleSubmit = () => {
+  emit('submit', { ...formData })
+}
+
+// Reset form when dialog opens
+watch(() => props.visible, (newVal) => {
+  if (newVal) {
+    formData.name = ''
+    formData.description = ''
+    formData.type = ''
+  }
+})
+</script>
+
+<style scoped>
+.create-dialog :deep(.el-dialog) {
+  border-radius: 20px;
+  background: linear-gradient(145deg, rgba(20, 20, 30, 0.98) 0%, rgba(15, 15, 25, 0.98) 100%);
+  border: 1px solid rgba(255, 255, 255, 0.08);
+  box-shadow:
+    0 25px 50px -12px rgba(0, 0, 0, 0.6),
+    0 0 40px rgba(0, 212, 255, 0.1),
+    inset 0 1px 0 rgba(255, 255, 255, 0.05);
+  overflow: hidden;
+}
+
+.create-dialog :deep(.el-dialog__header) {
+  padding: 0;
+  margin: 0;
+}
+
+.dialog-header {
+  position: relative;
+  padding: 24px 28px;
+  background: linear-gradient(135deg, rgba(0, 212, 255, 0.08) 0%, rgba(124, 58, 237, 0.08) 100%);
+  border-bottom: 1px solid rgba(255, 255, 255, 0.06);
+  overflow: hidden;
+}
+
+.header-glow {
+  position: absolute;
+  top: -50%;
+  left: -50%;
+  width: 200%;
+  height: 200%;
+  background: radial-gradient(circle at 30% 50%, rgba(0, 212, 255, 0.15) 0%, transparent 50%);
+  animation: headerGlow 4s ease-in-out infinite;
+}
+
+@keyframes headerGlow {
+  0%, 100% { opacity: 0.5; transform: scale(1); }
+  50% { opacity: 0.8; transform: scale(1.1); }
+}
+
+.header-content {
+  position: relative;
+  display: flex;
+  align-items: center;
+  gap: 16px;
+}
+
+.dialog-icon-new {
+  width: 52px;
+  height: 52px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: linear-gradient(135deg, rgba(0, 212, 255, 0.2) 0%, rgba(124, 58, 237, 0.2) 100%);
+  border: 1px solid rgba(0, 212, 255, 0.3);
+  border-radius: 14px;
+  color: var(--accent-primary);
+  animation: iconPulse 2s ease-in-out infinite;
+}
+
+@keyframes iconPulse {
+  0%, 100% { box-shadow: 0 0 0 0 rgba(0, 212, 255, 0.4); }
+  50% { box-shadow: 0 0 20px 5px rgba(0, 212, 255, 0.2); }
+}
+
+.header-text h3 {
+  font-size: 20px;
+  font-weight: 600;
+  color: var(--text-primary);
+  margin: 0 0 4px 0;
+}
+
+.header-text p {
+  font-size: 13px;
+  color: var(--text-tertiary);
+  margin: 0;
+}
+
+.close-btn {
+  position: absolute;
+  top: 16px;
+  right: 16px;
+  width: 32px;
+  height: 32px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: rgba(255, 255, 255, 0.05);
+  border: 1px solid rgba(255, 255, 255, 0.08);
+  border-radius: 8px;
+  color: var(--text-muted);
+  cursor: pointer;
+  transition: all 0.2s ease;
+}
+
+.close-btn:hover {
+  background: rgba(255, 255, 255, 0.1);
+  color: var(--text-primary);
+  border-color: rgba(255, 255, 255, 0.15);
+}
+
+/* Dialog Body */
+.create-dialog :deep(.el-dialog__body) {
+  padding: 28px;
+}
+
+.dialog-body {
+  display: flex;
+  flex-direction: column;
+  gap: 24px;
+}
+
+.form-card {
+  display: flex;
+  flex-direction: column;
+  gap: 20px;
+  padding: 24px;
+  background: rgba(255, 255, 255, 0.02);
+  border: 1px solid rgba(255, 255, 255, 0.05);
+  border-radius: 16px;
+}
+
+.input-group {
+  display: flex;
+  flex-direction: column;
+  gap: 8px;
+}
+
+.input-label {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  font-size: 13px;
+  font-weight: 500;
+  color: var(--text-secondary);
+}
+
+.label-icon {
+  color: var(--accent-primary);
+  font-size: 10px;
+}
+
+.custom-input :deep(.el-input__wrapper) {
+  background: rgba(0, 0, 0, 0.3);
+  border: 1px solid rgba(255, 255, 255, 0.08);
+  border-radius: 10px;
+  box-shadow: none;
+  padding: 4px 14px;
+  transition: all 0.25s ease;
+}
+
+.custom-input :deep(.el-input__wrapper:hover) {
+  border-color: rgba(0, 212, 255, 0.3);
+}
+
+.custom-input :deep(.el-input__wrapper.is-focus) {
+  border-color: var(--accent-primary);
+  box-shadow: 0 0 0 3px rgba(0, 212, 255, 0.15);
+}
+
+.custom-input :deep(.el-input__inner) {
+  color: var(--text-primary);
+}
+
+.custom-input :deep(.el-input__inner::placeholder) {
+  color: var(--text-muted);
+}
+
+.custom-input :deep(.el-textarea__inner) {
+  background: rgba(0, 0, 0, 0.3);
+  border: 1px solid rgba(255, 255, 255, 0.08);
+  border-radius: 10px;
+  box-shadow: none;
+  padding: 12px 14px;
+  color: var(--text-primary);
+  resize: none;
+  transition: all 0.25s ease;
+}
+
+.custom-input :deep(.el-textarea__inner::placeholder) {
+  color: var(--text-muted);
+}
+
+.custom-input :deep(.el-textarea__inner:hover) {
+  border-color: rgba(0, 212, 255, 0.3);
+}
+
+.custom-input :deep(.el-textarea__inner:focus) {
+  border-color: var(--accent-primary);
+  box-shadow: 0 0 0 3px rgba(0, 212, 255, 0.15);
+}
+
+/* Templates Section */
+.templates-section {
+  display: flex;
+  flex-direction: column;
+  gap: 12px;
+}
+
+.templates-label {
+  font-size: 12px;
+  font-weight: 500;
+  color: var(--text-muted);
+  text-transform: uppercase;
+  letter-spacing: 0.5px;
+}
+
+.templates-grid {
+  display: grid;
+  grid-template-columns: repeat(3, 1fr);
+  gap: 12px;
+}
+
+.template-card {
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  gap: 8px;
+  padding: 16px 12px;
+  background: rgba(255, 255, 255, 0.02);
+  border: 1px solid rgba(255, 255, 255, 0.06);
+  border-radius: 12px;
+  cursor: pointer;
+  transition: all 0.25s ease;
+}
+
+.template-card:hover {
+  background: rgba(0, 212, 255, 0.08);
+  border-color: rgba(0, 212, 255, 0.25);
+  transform: translateY(-2px);
+}
+
+.template-card.active {
+  background: rgba(0, 212, 255, 0.12);
+  border-color: var(--accent-primary);
+  box-shadow: 0 0 15px rgba(0, 212, 255, 0.2);
+}
+
+.template-card .el-icon {
+  font-size: 22px;
+  color: var(--accent-primary);
+}
+
+.template-card span {
+  font-size: 12px;
+  color: var(--text-secondary);
+}
+
+/* Dialog Footer */
+.create-dialog :deep(.el-dialog__footer) {
+  padding: 0;
+}
+
+.dialog-footer {
+  display: flex;
+  justify-content: center;
+  padding: 20px 28px;
+  background: rgba(0, 0, 0, 0.2);
+  border-top: 1px solid rgba(255, 255, 255, 0.05);
+}
+
+.btn-cancel {
+  padding: 10px 20px;
+  background: transparent;
+  border: 1px solid rgba(255, 255, 255, 0.1);
+  border-radius: 10px;
+  color: var(--text-secondary);
+  transition: all 0.2s ease;
+}
+
+.btn-cancel:hover {
+  background: rgba(255, 255, 255, 0.05);
+  border-color: rgba(255, 255, 255, 0.15);
+  color: var(--text-primary);
+}
+
+.btn-create {
+  width: 100%;
+  padding: 14px 32px;
+  background: linear-gradient(135deg, var(--accent-primary) 0%, #0891b2 100%);
+  border: none;
+  border-radius: 10px;
+  font-weight: 500;
+  font-size: 15px;
+  transition: all 0.25s ease;
+}
+
+.btn-create:hover {
+  transform: translateY(-1px);
+  box-shadow: 0 4px 15px rgba(0, 212, 255, 0.35);
+}
+</style>

frontend/src/shared/components/common/DeleteDialog.vue

@@ -0,0 +1,261 @@
+<template>
+  <el-dialog
+    :model-value="visible"
+    title=""
+    width="460px"
+    class="delete-dialog"
+    :show-close="false"
+    align-center
+    :close-on-click-modal="false"
+    @update:model-value="$emit('update:visible', $event)"
+  >
+    <template #header>
+      <div class="delete-dialog-header">
+        <div class="delete-header-top">
+          <div class="delete-icon-wrapper">
+            <el-icon size="26"><WarningFilled /></el-icon>
+          </div>
+          <div class="delete-kicker">Danger Zone</div>
+        </div>
+        <div class="delete-title-group">
+          <h3>{{ title }}</h3>
+          <p v-if="detailText" class="detail-text">{{ detailText }}</p>
+        </div>
+      </div>
+    </template>
+    <div class="delete-dialog-body">
+      <div class="delete-target-card">
+        <span class="target-label">目标对象</span>
+        <strong class="target-name">{{ itemName }}</strong>
+      </div>
+      <p class="warning-text">{{ warningText }}</p>
+    </div>
+    <template #footer>
+      <div class="delete-dialog-footer">
+        <el-button
+          @click="handleCancel"
+          class="btn-cancel-delete"
+        >
+          取消
+        </el-button>
+        <el-button
+          type="danger"
+          :loading="loading"
+          @click="handleConfirm"
+          class="btn-delete"
+        >
+          <el-icon v-if="!loading"><Delete /></el-icon>
+          {{ confirmText }}
+        </el-button>
+      </div>
+    </template>
+  </el-dialog>
+</template>
+
+<script setup>
+const props = defineProps({
+  visible: {
+    type: Boolean,
+    default: false
+  },
+  title: {
+    type: String,
+    default: '删除项目'
+  },
+  itemName: {
+    type: String,
+    default: ''
+  },
+  warningText: {
+    type: String,
+    default: '此操作不可恢复，所有相关数据将被永久删除'
+  },
+  detailText: {
+    type: String,
+    default: ''
+  },
+  confirmText: {
+    type: String,
+    default: '确认删除'
+  },
+  loading: {
+    type: Boolean,
+    default: false
+  }
+})
+
+const emit = defineEmits(['update:visible', 'confirm', 'cancel'])
+
+const handleConfirm = () => {
+  emit('confirm')
+}
+
+const handleCancel = () => {
+  emit('update:visible', false)
+  emit('cancel')
+}
+</script>
+
+<style scoped>
+.delete-dialog :deep(.el-dialog) {
+  background:
+    radial-gradient(circle at top left, rgba(239, 68, 68, 0.14), transparent 24%),
+    radial-gradient(circle at bottom right, rgba(251, 146, 60, 0.08), transparent 20%),
+    linear-gradient(180deg, #11151d 0%, #0d1118 100%);
+  border: 1px solid rgba(239, 68, 68, 0.18);
+  border-radius: 22px;
+  overflow: hidden;
+  box-shadow:
+    0 24px 80px rgba(0, 0, 0, 0.5),
+    inset 0 1px 0 rgba(255, 255, 255, 0.04);
+}
+
+.delete-dialog :deep(.el-dialog__header) {
+  padding: 0;
+  margin: 0;
+}
+
+.delete-dialog :deep(.el-dialog__body) {
+  padding: 0;
+}
+
+.delete-dialog :deep(.el-dialog__footer) {
+  padding: 0;
+}
+
+.delete-dialog-header {
+  padding: 24px 24px 18px;
+  display: flex;
+  flex-direction: column;
+  gap: 14px;
+  border-bottom: 1px solid rgba(255, 255, 255, 0.06);
+  background: linear-gradient(180deg, rgba(255, 255, 255, 0.03), transparent);
+}
+
+.delete-header-top {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+}
+
+.delete-icon-wrapper {
+  width: 48px;
+  height: 48px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: linear-gradient(135deg, rgba(239, 68, 68, 0.18), rgba(249, 115, 22, 0.08));
+  border: 1px solid rgba(239, 68, 68, 0.24);
+  border-radius: 14px;
+  color: #fb7185;
+  box-shadow: inset 0 1px 0 rgba(255, 255, 255, 0.05);
+}
+
+.delete-kicker {
+  font-size: 11px;
+  letter-spacing: 0.16em;
+  text-transform: uppercase;
+  color: rgba(251, 113, 133, 0.72);
+  font-family: 'JetBrains Mono', 'SF Mono', monospace;
+}
+
+.delete-title-group {
+  display: flex;
+  flex-direction: column;
+  gap: 6px;
+}
+
+.delete-dialog-header h3 {
+  margin: 0;
+  font-size: 22px;
+  font-weight: 600;
+  color: #f8fafc;
+  letter-spacing: -0.02em;
+}
+
+.detail-text {
+  margin: 0;
+  color: rgba(226, 232, 240, 0.72);
+  font-size: 13px;
+  line-height: 1.6;
+}
+
+.delete-dialog-body {
+  padding: 22px 24px 24px;
+}
+
+.delete-target-card {
+  display: flex;
+  flex-direction: column;
+  gap: 8px;
+  padding: 14px 16px;
+  border-radius: 16px;
+  background: rgba(255, 255, 255, 0.03);
+  border: 1px solid rgba(255, 255, 255, 0.06);
+  margin-bottom: 14px;
+}
+
+.target-label {
+  font-size: 11px;
+  letter-spacing: 0.12em;
+  text-transform: uppercase;
+  color: rgba(148, 163, 184, 0.65);
+  font-family: 'JetBrains Mono', 'SF Mono', monospace;
+}
+
+.target-name {
+  color: #f8fafc;
+  font-size: 15px;
+  font-weight: 600;
+  line-height: 1.5;
+  word-break: break-word;
+}
+
+.warning-text {
+  margin: 0;
+  font-size: 13px;
+  line-height: 1.7;
+  color: #fca5a5;
+}
+
+.delete-dialog-footer {
+  display: flex;
+  justify-content: flex-end;
+  gap: 12px;
+  padding: 18px 24px 24px;
+  background: linear-gradient(180deg, transparent, rgba(0, 0, 0, 0.14));
+  border-top: 1px solid rgba(255, 255, 255, 0.04);
+}
+
+.btn-cancel-delete {
+  min-width: 108px;
+  padding: 10px 20px;
+  background: rgba(255, 255, 255, 0.03);
+  border: 1px solid rgba(255, 255, 255, 0.08);
+  border-radius: 10px;
+  color: rgba(226, 232, 240, 0.8);
+  transition: all 0.2s ease;
+}
+
+.btn-cancel-delete:hover {
+  background: rgba(255, 255, 255, 0.06);
+  border-color: rgba(255, 255, 255, 0.16);
+  color: #fff;
+}
+
+.btn-delete {
+  min-width: 128px;
+  padding: 10px 20px;
+  background: linear-gradient(135deg, #ef4444 0%, #dc2626 55%, #b91c1c 100%);
+  border: none;
+  border-radius: 10px;
+  font-weight: 600;
+  transition: all 0.25s ease;
+  box-shadow: 0 10px 30px rgba(239, 68, 68, 0.18);
+}
+
+.btn-delete:hover {
+  transform: translateY(-1px);
+  box-shadow: 0 14px 34px rgba(239, 68, 68, 0.28);
+}
+</style>

frontend/src/shared/components/common/EmptyState.vue

@@ -0,0 +1,96 @@
+<template>
+  <div class="empty-state">
+    <div class="empty-illustration">
+      <div class="circle-1"></div>
+      <div class="circle-2"></div>
+      <div class="circle-3"></div>
+      <el-icon size="48"><component :is="icon" /></el-icon>
+    </div>
+    <h3>{{ title }}</h3>
+    <p>{{ description }}</p>
+    <el-button v-if="actionText" type="primary" @click="$emit('action')">
+      {{ actionText }}
+    </el-button>
+  </div>
+</template>
+
+<script setup>
+defineProps({
+  icon: {
+    type: Object,
+    default: () => null
+  },
+  title: {
+    type: String,
+    default: '暂无数据'
+  },
+  description: {
+    type: String,
+    default: '暂无相关内容'
+  },
+  actionText: {
+    type: String,
+    default: ''
+  }
+})
+
+defineEmits(['action'])
+</script>
+
+<style scoped>
+.empty-state {
+  grid-column: 1 / -1;
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  justify-content: center;
+  padding: 80px 40px;
+  background: var(--glass-bg);
+  backdrop-filter: blur(20px);
+  border: 1px dashed var(--border-default);
+  border-radius: var(--radius-xl);
+  text-align: center;
+}
+
+.empty-illustration {
+  position: relative;
+  width: 120px;
+  height: 120px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  margin-bottom: 24px;
+}
+
+.empty-illustration .el-icon {
+  position: relative;
+  z-index: 1;
+  color: var(--text-muted);
+}
+
+.circle-1, .circle-2, .circle-3 {
+  position: absolute;
+  border-radius: 50%;
+  border: 1px solid var(--border-default);
+}
+
+.circle-1 { width: 100%; height: 100%; animation: rotate 20s linear infinite; }
+.circle-2 { width: 70%; height: 70%; animation: rotate 15s linear infinite reverse; }
+.circle-3 { width: 40%; height: 40%; background: var(--bg-tertiary); }
+
+@keyframes rotate {
+  from { transform: rotate(0deg); }
+  to { transform: rotate(360deg); }
+}
+
+.empty-state h3 {
+  font-size: 20px;
+  margin-bottom: 8px;
+  color: var(--text-primary);
+}
+
+.empty-state p {
+  color: var(--text-tertiary);
+  margin-bottom: 24px;
+}
+</style>

frontend/src/shared/components/common/ProjectCard.vue

@@ -0,0 +1,211 @@
+<template>
+  <div
+    class="project-card"
+    :style="{ '--delay': delay }"
+    @click="$emit('click', project)"
+  >
+    <div class="card-glow"></div>
+    <button
+      v-if="showDelete"
+      class="card-delete-btn"
+      @click.stop="$emit('delete', project)"
+    >
+      <el-icon><Delete /></el-icon>
+    </button>
+    <div class="card-header">
+      <div class="card-avatar">
+        <el-icon><component :is="projectIcon" /></el-icon>
+      </div>
+    </div>
+    <h3 class="card-title">{{ project.name }}</h3>
+    <p class="card-desc">{{ project.description || '暂无描述' }}</p>
+    <div class="card-footer">
+      <span class="card-date">
+        <el-icon><Calendar /></el-icon>
+        {{ formattedDate }}
+      </span>
+      <div class="card-status">
+        <span class="status-dot"></span>
+        {{ statusText }}
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { computed } from 'vue'
+import { Folder, ChatDotRound, Document, Connection } from '@element-plus/icons-vue'
+
+const props = defineProps({
+  project: {
+    type: Object,
+    required: true
+  },
+  index: {
+    type: Number,
+    default: 0
+  },
+  showDelete: {
+    type: Boolean,
+    default: true
+  },
+  statusText: {
+    type: String,
+    default: '活跃'
+  }
+})
+
+defineEmits(['click', 'delete'])
+
+const delay = computed(() => `${props.index * 0.1}s`)
+
+const projectIcon = computed(() => {
+  const type = props.project.type
+  if (type === 'qa') return ChatDotRound
+  if (type === 'table') return Document
+  if (type === 'database') return Connection
+  return Folder
+})
+
+const formattedDate = computed(() => {
+  if (!props.project.created_at) return ''
+  const d = new Date(props.project.created_at)
+  return d.toLocaleDateString('zh-CN', { month: 'short', day: 'numeric', year: 'numeric' })
+})
+</script>
+
+<style scoped>
+.project-card {
+  position: relative;
+  padding: 24px;
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-subtle);
+  border-radius: var(--radius-lg);
+  cursor: pointer;
+  transition: all var(--transition-base);
+  overflow: hidden;
+  animation: cardFadeIn 0.5s ease backwards;
+  animation-delay: var(--delay);
+}
+
+@keyframes cardFadeIn {
+  from { opacity: 0; transform: translateY(20px); }
+  to { opacity: 1; transform: translateY(0); }
+}
+
+.project-card:hover {
+  border-color: var(--accent-primary);
+  transform: translateY(-4px);
+}
+
+.card-glow {
+  position: absolute;
+  top: 0;
+  left: 0;
+  right: 0;
+  height: 1px;
+  background: linear-gradient(90deg, transparent, var(--accent-primary), transparent);
+  opacity: 0;
+  transition: opacity var(--transition-base);
+}
+
+.project-card:hover .card-glow { opacity: 1; }
+
+.card-delete-btn {
+  position: absolute;
+  top: 12px;
+  right: 12px;
+  width: 32px;
+  height: 32px;
+  border-radius: 50%;
+  background: var(--danger);
+  border: none;
+  color: white;
+  cursor: pointer;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  opacity: 0;
+  transform: scale(0.8);
+  transition: all 0.2s ease;
+  z-index: 10;
+}
+
+.card-delete-btn:hover {
+  transform: scale(1.1);
+  background: #dc2626;
+}
+
+.project-card:hover .card-delete-btn {
+  opacity: 1;
+  transform: scale(1);
+}
+
+.card-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: flex-start;
+  margin-bottom: 16px;
+}
+
+.card-avatar {
+  width: 44px;
+  height: 44px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: var(--accent-primary-muted);
+  border-radius: var(--radius-md);
+  font-size: 20px;
+  color: var(--accent-primary);
+}
+
+.card-title {
+  font-size: 17px;
+  font-weight: 600;
+  margin-bottom: 8px;
+  color: var(--text-primary);
+}
+
+.card-desc {
+  font-size: 14px;
+  color: var(--text-tertiary);
+  line-height: 1.5;
+  margin-bottom: 20px;
+  display: -webkit-box;
+  -webkit-line-clamp: 2;
+  -webkit-box-orient: vertical;
+  overflow: hidden;
+}
+
+.card-footer {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  padding-top: 16px;
+  border-top: 1px solid var(--border-subtle);
+}
+
+.card-date {
+  display: flex;
+  align-items: center;
+  gap: 6px;
+  font-size: 13px;
+  color: var(--text-muted);
+}
+
+.card-status {
+  display: flex;
+  align-items: center;
+  gap: 6px;
+  font-size: 13px;
+  color: var(--success);
+}
+
+.status-dot {
+  width: 6px;
+  height: 6px;
+  background: var(--success);
+  border-radius: 50%;
+}
+</style>

frontend/src/shared/composables/index.ts

@@ -0,0 +1,6 @@
+/**
+ * Composables - 可复用业务逻辑
+ */
+export * from './useFormatters'
+export * from './useProjects'
+export * from './useModels'

frontend/src/shared/composables/useFormatters.ts

@@ -0,0 +1,71 @@
+/**
+ * 格式化工具函数
+ */
+
+/**
+ * 格式化文件大小
+ */
+export function formatSize(bytes: number): string {
+  if (bytes === 0) return '0 B'
+  const k = 1024
+  const sizes = ['B', 'KB', 'MB', 'GB', 'TB']
+  const i = Math.floor(Math.log(bytes) / Math.log(k))
+  return parseFloat((bytes / Math.pow(k, i)).toFixed(2)) + ' ' + sizes[i]
+}
+
+/**
+ * 格式化日期
+ */
+export function formatDate(date: string | Date): string {
+  const d = new Date(date)
+  const year = d.getFullYear()
+  const month = String(d.getMonth() + 1).padStart(2, '0')
+  const day = String(d.getDate()).padStart(2, '0')
+  return `${year}-${month}-${day}`
+}
+
+/**
+ * 格式化日期时间
+ */
+export function formatDateTime(date: string | Date): string {
+  const d = new Date(date)
+  const year = d.getFullYear()
+  const month = String(d.getMonth() + 1).padStart(2, '0')
+  const day = String(d.getDate()).padStart(2, '0')
+  const hours = String(d.getHours()).padStart(2, '0')
+  const minutes = String(d.getMinutes()).padStart(2, '0')
+  return `${year}-${month}-${day} ${hours}:${minutes}`
+}
+
+/**
+ * 格式化相对时间
+ */
+export function formatRelativeTime(date: string | Date): string {
+  const now = new Date()
+  const d = new Date(date)
+  const diff = now.getTime() - d.getTime()
+
+  const seconds = Math.floor(diff / 1000)
+  const minutes = Math.floor(seconds / 60)
+  const hours = Math.floor(minutes / 60)
+  const days = Math.floor(hours / 24)
+
+  if (days > 7) {
+    return formatDate(date)
+  } else if (days > 0) {
+    return `${days} 天前`
+  } else if (hours > 0) {
+    return `${hours} 小时前`
+  } else if (minutes > 0) {
+    return `${minutes} 分钟前`
+  } else {
+    return '刚刚'
+  }
+}
+
+/**
+ * 格式化数字（千分位）
+ */
+export function formatNumber(num: number): string {
+  return num.toLocaleString('zh-CN')
+}

frontend/src/shared/composables/useModels.ts

@@ -0,0 +1,112 @@
+/**
+ * 模型相关业务逻辑
+ */
+import { ref } from 'vue'
+import { modelApi } from '@/core/api'
+import type { Model } from '@/shared/types'
+import { ElMessage } from 'element-plus'
+
+export function useModels() {
+  const loading = ref(false)
+  const models = ref<Model[]>([])
+
+  /**
+   * 获取模型列表
+   */
+  const fetchModels = async () => {
+    loading.value = true
+    try {
+      const res = await modelApi.list()
+      // 处理两种响应格式
+      if (Array.isArray(res)) {
+        models.value = res
+      } else if (res?.data && Array.isArray(res.data)) {
+        models.value = res.data
+      } else if (res?.results && Array.isArray(res.results)) {
+        models.value = res.results
+      } else {
+        models.value = []
+      }
+    } catch (error: any) {
+      console.error('获取模型列表失败:', error)
+      ElMessage.error('获取模型列表失败')
+      models.value = []
+    } finally {
+      loading.value = false
+    }
+  }
+
+  /**
+   * 添加模型
+   */
+  const addModel = async (data: Partial<Model>): Promise<boolean> => {
+    try {
+      await modelApi.create(data)
+      ElMessage.success('添加成功')
+      await fetchModels()
+      return true
+    } catch (error: any) {
+      console.error('添加模型失败:', error)
+      ElMessage.error(error?.message || '添加模型失败')
+      return false
+    }
+  }
+
+  /**
+   * 更新模型
+   */
+  const updateModel = async (id: number, data: Partial<Model>): Promise<boolean> => {
+    try {
+      await modelApi.update(id, data)
+      ElMessage.success('更新成功')
+      await fetchModels()
+      return true
+    } catch (error: any) {
+      console.error('更新模型失败:', error)
+      ElMessage.error(error?.message || '更新模型失败')
+      return false
+    }
+  }
+
+  /**
+   * 删除模型
+   */
+  const deleteModel = async (id: number): Promise<boolean> => {
+    try {
+      await modelApi.delete(id)
+      ElMessage.success('删除成功')
+      await fetchModels()
+      return true
+    } catch (error: any) {
+      console.error('删除模型失败:', error)
+      ElMessage.error(error?.message || '删除模型失败')
+      return false
+    }
+  }
+
+  /**
+   * 设置默认模型
+   */
+  const setDefaultModel = async (id: number): Promise<boolean> => {
+    try {
+      await modelApi.setDefault(id)
+      ElMessage.success('设置成功')
+      await fetchModels()
+      return true
+    } catch (error: any) {
+      console.error('设置默认模型失败:', error)
+      ElMessage.error(error?.message || '设置默认模型失败')
+      return false
+    }
+  }
+
+  return {
+    loading,
+    models,
+    fetchModels,
+    addModel,
+    updateModel,
+    deleteModel,
+    setDefaultModel
+  }
+}

frontend/src/shared/composables/useProjects.ts

@@ -0,0 +1,98 @@
+/**
+ * 项目相关业务逻辑
+ */
+import { ref } from 'vue'
+import { projectApi } from '@/core/api'
+import type { Project, ProjectCreate } from '@/shared/types'
+import { ElMessage } from 'element-plus'
+
+export function useProjects() {
+  const loading = ref(false)
+  const projects = ref<Project[]>([])
+
+  /**
+   * 获取项目列表
+   */
+  const fetchProjects = async () => {
+    loading.value = true
+    try {
+      const res = await projectApi.list()
+      // 处理两种响应格式
+      if (Array.isArray(res)) {
+        projects.value = res
+      } else if (res?.data && Array.isArray(res.data)) {
+        projects.value = res.data
+      } else if (res?.results && Array.isArray(res.results)) {
+        projects.value = res.results
+      } else {
+        projects.value = []
+      }
+    } catch (error: any) {
+      console.error('获取项目列表失败:', error)
+      ElMessage.error('获取项目列表失败')
+      projects.value = []
+    } finally {
+      loading.value = false
+    }
+  }
+
+  /**
+   * 创建项目
+   */
+  const createProject = async (data: ProjectCreate): Promise<Project | null> => {
+    try {
+      const res = await projectApi.create(data)
+      ElMessage.success('创建成功')
+      await fetchProjects()
+      return res
+    } catch (error: any) {
+      console.error('创建项目失败:', error)
+      ElMessage.error(error?.message || '创建项目失败')
+      return null
+    }
+  }
+
+  /**
+   * 删除项目
+   */
+  const deleteProject = async (id: number): Promise<boolean> => {
+    try {
+      await projectApi.delete(id)
+      ElMessage.success('删除成功')
+      await fetchProjects()
+      return true
+    } catch (error: any) {
+      console.error('删除项目失败:', error)
+      ElMessage.error(error?.message || '删除项目失败')
+      return false
+    }
+  }
+
+  /**
+   * 获取项目详情
+   */
+  const fetchProject = async (id: number): Promise<Project | null> => {
+    try {
+      const res = await projectApi.get(id)
+      if (res && typeof res === 'object' && 'id' in res) {
+        return res as Project
+      } else if (res?.data) {
+        return res.data as Project
+      }
+      return null
+    } catch (error: any) {
+      console.error('获取项目详情失败:', error)
+      ElMessage.error('获取项目详情失败')
+      return null
+    }
+  }
+
+  return {
+    loading,
+    projects,
+    fetchProjects,
+    createProject,
+    deleteProject,
+    fetchProject
+  }
+}

frontend/src/shared/types/api.d.ts

@@ -0,0 +1,32 @@
+/**
+ * API Response Types
+ */
+
+// Base API response wrapper
+export interface ApiResponse<T = any> {
+  success: boolean
+  message: string
+  data: T
+  error: string | null
+  timestamp: string
+}
+
+// Paginated response
+export interface PaginatedResponse<T = any> extends ApiResponse<T> {
+  page?: number
+  page_size?: number
+  total?: number
+}
+
+// List items wrapper
+export interface ListResponse<T> {
+  items: T[]
+  total: number
+  page: number
+  page_size: number
+}
+
+// Simple ID response
+export interface IdResponse {
+  id: string
+}

frontend/src/shared/types/common.d.ts

@@ -0,0 +1,60 @@
+/**
+ * Common Types
+ */
+
+// File types
+export interface FileItem {
+  id: string
+  filename: string
+  file_type: string
+  size?: number
+  status: string
+  created_at: string
+  updated_at: string
+}
+
+// Chunk types
+export interface Chunk {
+  id: string
+  name?: string
+  content: string
+  summary?: string
+  word_count?: number
+  file_id?: string
+  created_at: string
+  updated_at: string
+}
+
+// Question types
+export interface Question {
+  id: string
+  content: string
+  answer?: string
+  question_type?: string
+  chunk_id?: string
+  source: string
+  created_at: string
+  updated_at: string
+}
+
+// Dataset types
+export interface Dataset {
+  id: string
+  name: string
+  description?: string
+  dataset_type?: string
+  question_count?: number
+  created_at: string
+  updated_at: string
+}
+
+// Dialog props
+export interface DialogProps {
+  visible: boolean
+  loading?: boolean
+}
+
+export interface DeleteDialogProps extends DialogProps {
+  itemName?: string
+  itemType?: string
+}

frontend/src/shared/types/index.ts

@@ -0,0 +1,7 @@
+/**
+ * Type exports
+ */
+export * from './api'
+export * from './project'
+export * from './model'
+export * from './common'

frontend/src/shared/types/model.d.ts

@@ -0,0 +1,47 @@
+/**
+ * Model Configuration Types
+ */
+
+export interface Model {
+  id: string
+  provider: ModelProvider
+  model_type: ModelType
+  model_name: string
+  api_key?: string
+  api_base?: string
+  is_default: 'true' | 'false'
+  connection_status?: 'untested' | 'connected' | 'disconnected'
+  created_at?: string
+  updated_at?: string
+}
+
+export interface ModelConfig {
+  id: string
+  provider: ModelProvider
+  model_type: ModelType
+  model_name: string
+  api_key?: string
+  api_base?: string
+  is_default: 'true' | 'false'
+  connection_status?: 'untested' | 'connected' | 'disconnected'
+  created_at?: string
+  updated_at?: string
+}
+
+export type ModelProvider = 'minimax' | 'glm' | 'openai' | 'ali'
+export type ModelType = 'chat' | 'vlm' | 'embedding' | 'rerank'
+
+export interface ModelCreate {
+  provider: ModelProvider
+  model_type: ModelType
+  model_name: string
+  api_key: string
+  api_base?: string
+  is_default: boolean
+}
+
+export interface ProviderOption {
+  value: ModelProvider
+  label: string
+  abbr: string
+}

frontend/src/shared/types/project.d.ts

@@ -0,0 +1,24 @@
+/**
+ * Project Types
+ */
+
+export interface Project {
+  id: string
+  name: string
+  description?: string
+  type: string
+  created_at: string
+  updated_at: string
+}
+
+export interface ProjectCreate {
+  name: string
+  description: string
+  type: string
+}
+
+export interface ProjectUpdate {
+  name?: string
+  description?: string
+  type?: string
+}