ns2.0/backend/app/api/apps.py

from fastapi import APIRouter, Depends, HTTPException, Query
from sqlalchemy.ext.asyncio import AsyncSession
from sqlalchemy import select, func, and_, or_
from datetime import datetime, timedelta
from typing import Optional
from app.database import get_db
from app.models import AppInfo, AppMetrics, AppRating
from app.schemas import ApiResponse
from app.crawler.huawei_api import HuaweiAPI
from app.crawler.data_processor import DataProcessor

router = APIRouter(prefix="/apps", tags=["应用"])

@router.get("/fetch/{pkg_name}")
async def fetch_app_by_pkg_name(
    pkg_name: str,
    db: AsyncSession = Depends(get_db)
):
    """通过包名从华为API获取应用信息并保存"""
    api = HuaweiAPI()
    try:
        # 从华为API获取数据
        print(f"正在获取应用信息: {pkg_name}")
        app_data = await api.get_app_info(pkg_name=pkg_name)
        
        # 获取评分数据
        rating_data = await api.get_app_rating(app_data['appId'])
        
        # 保存到数据库
        processor = DataProcessor(db)
        new_info, new_metric, new_rating = await processor.save_app_data(
            app_data, rating_data
        )
        
        return ApiResponse(
            success=True,
            data={
                "app_id": app_data['appId'],
                "name": app_data['name'],
                "pkg_name": app_data['pkgName'],
                "new_info": new_info,
                "new_metric": new_metric,
                "new_rating": new_rating,
                "message": "应用信息获取成功"
            }
        )
    
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"获取应用信息失败: {str(e)}")
    finally:
        await api.close()

@router.get("/search")
async def search_apps(
    q: str = Query(..., min_length=1),
    page: int = Query(1, ge=1),
    page_size: int = Query(20, le=100),
    db: AsyncSession = Depends(get_db)
):
    """搜索应用"""
    subquery = (
        select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
        .group_by(AppMetrics.app_id)
        .subquery()
    )
    
    query = (
        select(AppInfo, AppMetrics, AppRating)
        .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
        .outerjoin(AppRating, AppInfo.app_id == AppRating.app_id)
        .join(subquery, and_(
            AppMetrics.app_id == subquery.c.app_id,
            AppMetrics.created_at == subquery.c.max_created_at
        ))
        .where(or_(
            AppInfo.name.like(f"%{q}%"),
            AppInfo.pkg_name.like(f"%{q}%"),
            AppInfo.developer_name.like(f"%{q}%")
        ))
        .order_by(AppMetrics.download_count.desc())
    )
    
    count_query = select(func.count(AppInfo.app_id)).where(or_(
        AppInfo.name.like(f"%{q}%"),
        AppInfo.pkg_name.like(f"%{q}%"),
        AppInfo.developer_name.like(f"%{q}%")
    ))
    
    total_result = await db.execute(count_query)
    total = total_result.scalar()
    
    offset = (page - 1) * page_size
    query = query.offset(offset).limit(page_size)
    
    result = await db.execute(query)
    rows = result.all()
    
    data = [{
        "app_id": row[0].app_id,
        "name": row[0].name,
        "pkg_name": row[0].pkg_name,
        "developer_name": row[0].developer_name,
        "kind_name": row[0].kind_name,
        "icon_url": row[0].icon_url,
        "brief_desc": row[0].brief_desc,
        "download_count": row[1].download_count if len(row) > 1 else 0,
        "version": row[1].version if len(row) > 1 else "",
        "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0,
        "listed_at": row[0].listed_at.isoformat()
    } for row in rows]
    
    return ApiResponse(success=True, data=data, total=total, limit=page_size)

@router.get("/category/{category}")
async def get_apps_by_category(
    category: str,
    page: int = Query(1, ge=1),
    page_size: int = Query(20, le=100),
    db: AsyncSession = Depends(get_db)
):
    """按分类获取应用"""
    subquery = (
        select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
        .group_by(AppMetrics.app_id)
        .subquery()
    )
    
    # 构建基础查询
    query = (
        select(AppInfo, AppMetrics, AppRating)
        .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
        .outerjoin(AppRating, AppInfo.app_id == AppRating.app_id)
        .join(subquery, and_(
            AppMetrics.app_id == subquery.c.app_id,
            AppMetrics.created_at == subquery.c.max_created_at
        ))
    )
    
    # 如果是元服务分类，只显示元服务（packing_type = 1）
    if category == "元服务":
        query = query.where(AppInfo.packing_type == 1)
    else:
        # 其他分类排除元服务，并按kind_name筛选
        query = query.where(and_(
            AppInfo.kind_name == category,
            or_(AppInfo.packing_type != 1, AppInfo.packing_type.is_(None))
        ))
    
    query = query.order_by(AppMetrics.download_count.desc())
)
    
    count_query = select(func.count(AppInfo.app_id)).where(AppInfo.kind_name == category)
    total_result = await db.execute(count_query)
    total = total_result.scalar()
    
    offset = (page - 1) * page_size
    query = query.offset(offset).limit(page_size)
    
    result = await db.execute(query)
    rows = result.all()
    
    data = [{
        "app_id": row[0].app_id,
        "name": row[0].name,
        "pkg_name": row[0].pkg_name,
        "developer_name": row[0].developer_name,
        "kind_name": row[0].kind_name,
        "icon_url": row[0].icon_url,
        "brief_desc": row[0].brief_desc,
        "download_count": row[1].download_count if len(row) > 1 else 0,
        "version": row[1].version if len(row) > 1 else "",
        "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0,
        "listed_at": row[0].listed_at.isoformat()
    } for row in rows]
    
    return ApiResponse(success=True, data=data, total=total, limit=page_size)

@router.get("/categories")
async def get_categories(db: AsyncSession = Depends(get_db)):
    """获取所有分类"""
    # 获取元服务数量
    atomic_service_result = await db.execute(
        select(func.count(AppInfo.app_id))
        .where(AppInfo.packing_type == 1)
    )
    atomic_service_count = atomic_service_result.scalar()
    
    # 获取其他分类（排除元服务）
    result = await db.execute(
        select(AppInfo.kind_name, func.count(AppInfo.app_id).label('count'))
        .where(or_(AppInfo.packing_type != 1, AppInfo.packing_type.is_(None)))
        .group_by(AppInfo.kind_name)
        .order_by(func.count(AppInfo.app_id).desc())
    )
    rows = result.all()
    
    data = []
    
    # 如果有元服务，添加到列表首位
    if atomic_service_count > 0:
        data.append({"name": "元服务", "count": atomic_service_count})
    
    # 添加其他分类
    data.extend([{"name": row[0], "count": row[1]} for row in rows])
    
    return ApiResponse(success=True, data=data)

@router.get("/by-date")
async def get_apps_by_date(
    date: str = Query(..., description="日期格式: YYYY-MM-DD"),
    page_size: int = Query(100, le=100),
    db: AsyncSession = Depends(get_db)
):
    """获取指定日期上架的应用"""
    try:
        from datetime import datetime, time
        
        # 解析日期字符串
        target_date = datetime.strptime(date, '%Y-%m-%d')
        date_start = datetime.combine(target_date, time.min)
        date_end = datetime.combine(target_date, time.max)
        
        # 获取最新的指标记录
        subquery = (
            select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
            .group_by(AppMetrics.app_id)
            .subquery()
        )
        
        # 查询指定日期上架的应用
        query = (
            select(AppInfo, AppMetrics, AppRating)
            .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
            .outerjoin(AppRating, AppInfo.app_id == AppRating.app_id)
            .join(subquery, and_(
                AppMetrics.app_id == subquery.c.app_id,
                AppMetrics.created_at == subquery.c.max_created_at
            ))
            .where(and_(
                AppInfo.listed_at >= date_start,
                AppInfo.listed_at <= date_end
            ))
            .order_by(AppInfo.listed_at.desc())
            .limit(page_size)
        )
        
        result = await db.execute(query)
        rows = result.all()
        
        data = [{
            "app_id": row[0].app_id,
            "name": row[0].name,
            "pkg_name": row[0].pkg_name,
            "developer_name": row[0].developer_name,
            "kind_name": row[0].kind_name,
            "icon_url": row[0].icon_url,
            "brief_desc": row[0].brief_desc,
            "download_count": row[1].download_count if len(row) > 1 and row[1] else 0,
            "version": row[1].version if len(row) > 1 and row[1] else "",
            "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0.0,
            "total_rating_count": row[2].total_rating_count if len(row) > 2 and row[2] else 0,
            "listed_at": row[0].listed_at.isoformat() if row[0].listed_at else ""
        } for row in rows]
        
        return ApiResponse(success=True, data=data, total=len(data))
    except ValueError as e:
        raise HTTPException(status_code=400, detail=f"日期格式错误: {str(e)}")
    except Exception as e:
        print(f"Error in get_apps_by_date: {e}")
        import traceback
        traceback.print_exc()
        return ApiResponse(success=True, data=[], total=0)

@router.get("/today")
async def get_today_apps(
    page_size: int = Query(100, le=100),
    db: AsyncSession = Depends(get_db)
):
    """获取今日上架应用（根据 listed_at 字段判断是否为今天上架）"""
    try:
        # 获取今天的日期范围（00:00:00 到 23:59:59）
        from datetime import datetime, time
        today_start = datetime.combine(datetime.today(), time.min)
        today_end = datetime.combine(datetime.today(), time.max)
        
        # 获取最新的指标记录
        subquery = (
            select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
            .group_by(AppMetrics.app_id)
            .subquery()
        )
        
        # 查询今天上架的应用（根据 listed_at 字段）
        query = (
            select(AppInfo, AppMetrics, AppRating)
            .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
            .outerjoin(AppRating, AppInfo.app_id == AppRating.app_id)
            .join(subquery, and_(
                AppMetrics.app_id == subquery.c.app_id,
                AppMetrics.created_at == subquery.c.max_created_at
            ))
            .where(and_(
                AppInfo.listed_at >= today_start,
                AppInfo.listed_at <= today_end
            ))
            .order_by(AppInfo.listed_at.desc())
            .limit(page_size)
        )
        
        result = await db.execute(query)
        rows = result.all()
        
        data = [{
            "app_id": row[0].app_id,
            "name": row[0].name,
            "pkg_name": row[0].pkg_name,
            "developer_name": row[0].developer_name,
            "kind_name": row[0].kind_name,
            "icon_url": row[0].icon_url,
            "brief_desc": row[0].brief_desc,
            "download_count": row[1].download_count if len(row) > 1 and row[1] else 0,
            "version": row[1].version if len(row) > 1 and row[1] else "",
            "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0.0,
            "total_rating_count": row[2].total_rating_count if len(row) > 2 and row[2] else 0,
            "listed_at": row[0].listed_at.isoformat() if row[0].listed_at else ""
        } for row in rows]
        
        return ApiResponse(success=True, data=data, total=len(data))
    except Exception as e:
        print(f"Error in get_today_apps: {e}")
        import traceback
        traceback.print_exc()
        # 返回空列表而不是抛出错误
        return ApiResponse(success=True, data=[], total=0)

@router.get("/top-downloads")
async def get_top_downloads(
    limit: int = Query(100, le=100),
    db: AsyncSession = Depends(get_db)
):
    """热门应用Top100"""
    # 最新的指标记录
    subquery_metric = (
        select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
        .group_by(AppMetrics.app_id)
        .subquery()
    )
    
    # 最新的评分记录
    subquery_rating = (
        select(AppRating.app_id, func.max(AppRating.created_at).label('max_rating_created_at'))
        .group_by(AppRating.app_id)
        .subquery()
    )
    
    query = (
        select(AppInfo, AppMetrics, AppRating)
        .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
        .join(subquery_metric, and_(
            AppMetrics.app_id == subquery_metric.c.app_id,
            AppMetrics.created_at == subquery_metric.c.max_created_at
        ))
        .outerjoin(subquery_rating, AppInfo.app_id == subquery_rating.c.app_id)
        .outerjoin(AppRating, and_(
            AppInfo.app_id == AppRating.app_id,
            AppRating.created_at == subquery_rating.c.max_rating_created_at
        ))
        .order_by(AppMetrics.download_count.desc())
        .limit(limit)
    )
    
    result = await db.execute(query)
    rows = result.all()
    
    data = [{
        "app_id": row[0].app_id,
        "name": row[0].name,
        "pkg_name": row[0].pkg_name,
        "developer_name": row[0].developer_name,
        "kind_name": row[0].kind_name,
        "icon_url": row[0].icon_url,
        "brief_desc": row[0].brief_desc,
        "download_count": row[1].download_count if len(row) > 1 else 0,
        "version": row[1].version if len(row) > 1 else "",
        "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0
    } for row in rows]
    
    return ApiResponse(success=True, data=data, total=len(data))

@router.get("/top-ratings")
async def get_top_ratings(
    limit: int = Query(100, le=100),
    db: AsyncSession = Depends(get_db)
):
    """评分Top100"""
    subquery_metric = (
        select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
        .group_by(AppMetrics.app_id)
        .subquery()
    )
    
    subquery_rating = (
        select(AppRating.app_id, func.max(AppRating.created_at).label('max_created_at'))
        .group_by(AppRating.app_id)
        .subquery()
    )
    
    query = (
        select(AppInfo, AppMetrics, AppRating)
        .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
        .join(AppRating, AppInfo.app_id == AppRating.app_id)
        .join(subquery_metric, and_(
            AppMetrics.app_id == subquery_metric.c.app_id,
            AppMetrics.created_at == subquery_metric.c.max_created_at
        ))
        .join(subquery_rating, and_(
            AppRating.app_id == subquery_rating.c.app_id,
            AppRating.created_at == subquery_rating.c.max_created_at
        ))
        .where(AppRating.total_rating_count >= 100)
        .order_by(AppRating.average_rating.desc())
        .limit(limit)
    )
    
    result = await db.execute(query)
    rows = result.all()
    
    data = [{
        "app_id": row[0].app_id,
        "name": row[0].name,
        "pkg_name": row[0].pkg_name,
        "developer_name": row[0].developer_name,
        "kind_name": row[0].kind_name,
        "icon_url": row[0].icon_url,
        "brief_desc": row[0].brief_desc,
        "download_count": row[1].download_count if len(row) > 1 else 0,
        "version": row[1].version if len(row) > 1 else "",
        "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0,
        "total_rating_count": row[2].total_rating_count if len(row) > 2 and row[2] else 0
    } for row in rows]
    
    return ApiResponse(success=True, data=data, total=len(data))

@router.get("/{app_id}")
async def get_app_detail(app_id: str, db: AsyncSession = Depends(get_db)):
    """获取应用详情"""
    subquery = (
        select(AppMetrics.app_id, func.max(AppMetrics.created_at).label('max_created_at'))
        .where(AppMetrics.app_id == app_id)
        .group_by(AppMetrics.app_id)
        .subquery()
    )
    
    query = (
        select(AppInfo, AppMetrics, AppRating)
        .join(AppMetrics, AppInfo.app_id == AppMetrics.app_id)
        .outerjoin(AppRating, AppInfo.app_id == AppRating.app_id)
        .join(subquery, and_(
            AppMetrics.app_id == subquery.c.app_id,
            AppMetrics.created_at == subquery.c.max_created_at
        ))
        .where(AppInfo.app_id == app_id)
    )
    
    result = await db.execute(query)
    row = result.first()
    
    if not row:
        raise HTTPException(status_code=404, detail="应用不存在")
    
    data = {
        # 基本信息
        "app_id": row[0].app_id,
        "name": row[0].name,
        "pkg_name": row[0].pkg_name,
        
        # 开发者信息
        "developer_name": row[0].developer_name,
        "dev_id": row[0].dev_id,
        "supplier": row[0].supplier,
        
        # 分类信息
        "kind_name": row[0].kind_name,
        "kind_id": row[0].kind_id,
        "tag_name": row[0].tag_name,
        
        # 展示信息
        "icon_url": row[0].icon_url,
        "brief_desc": row[0].brief_desc,
        "description": row[0].description,
        
        # 隐私和政策
        "privacy_url": row[0].privacy_url,
        
        # 价格和支付
        "is_pay": row[0].is_pay,
        "price": row[0].price,
        
        # 时间信息
        "listed_at": row[0].listed_at.isoformat(),
        
        # 设备支持
        "main_device_codes": row[0].main_device_codes or [],
        
        # SDK信息
        "target_sdk": row[0].target_sdk,
        "min_sdk": row[0].min_sdk,
        "compile_sdk_version": row[0].compile_sdk_version,
        "min_hmos_api_level": row[0].min_hmos_api_level,
        "api_release_type": row[0].api_release_type,
        
        # 其他信息
        "ctype": row[0].ctype,
        "app_level": row[0].app_level,
        "packing_type": row[0].packing_type,
        
        # 版本和指标信息
        "download_count": row[1].download_count if len(row) > 1 else 0,
        "version": row[1].version if len(row) > 1 else "",
        "size_bytes": row[1].size_bytes if len(row) > 1 else 0,
        
        # 评分信息
        "average_rating": float(row[2].average_rating) if len(row) > 2 and row[2] else 0,
        "total_rating_count": row[2].total_rating_count if len(row) > 2 and row[2] else 0,
        "star_1_count": row[2].star_1_count if len(row) > 2 and row[2] else 0,
        "star_2_count": row[2].star_2_count if len(row) > 2 and row[2] else 0,
        "star_3_count": row[2].star_3_count if len(row) > 2 and row[2] else 0,
        "star_4_count": row[2].star_4_count if len(row) > 2 and row[2] else 0,
        "star_5_count": row[2].star_5_count if len(row) > 2 and row[2] else 0
    }
    
    return ApiResponse(success=True, data=data)