app_03_pregnancy/backend/app/routers/report.py

import logging
from datetime import datetime, timezone, date

from fastapi import APIRouter, Depends, HTTPException, UploadFile, File

from app.middleware.auth import get_current_user
from app.models.report import Report, DataItem
from app.models.user import User
from app.schemas.report import (
    ReportVerifyRequest,
    ReportResponse,
    ReportListItem,
    DataItemSchema,
)
from app.services.storage import save_upload_file
from app.services.ocr import ocr_recognize
from app.utils.helpers import parse_object_id, mask_name

logger = logging.getLogger(__name__)

router = APIRouter()

ALLOWED_CONTENT_TYPES = {"image/jpeg", "image/png", "image/webp"}
MAX_FILE_SIZE = 10 * 1024 * 1024  # 10MB


@router.post("/upload")
async def upload_report_image(
    file: UploadFile = File(...),
    user: User = Depends(get_current_user),
):
    """上传报告图片并进行 OCR 识别"""
    # Validate file type
    if file.content_type not in ALLOWED_CONTENT_TYPES:
        raise HTTPException(
            status_code=400,
            detail=f"不支持的文件类型: {file.content_type}，请上传 JPG/PNG/WebP 图片",
        )

    # Validate file size (read content and check)
    content = await file.read()
    if len(content) > MAX_FILE_SIZE:
        raise HTTPException(status_code=400, detail="文件大小不能超过 10MB")
    # Reset file position for storage
    await file.seek(0)

    url = await save_upload_file(file, sub_dir="reports")

    # OCR recognition
    try:
        ocr_result = await ocr_recognize(url)
    except FileNotFoundError:
        raise HTTPException(status_code=500, detail="文件保存异常，请重试")
    except RuntimeError as e:
        # OCR 服务已捕获并转换为友好消息
        logger.warning("OCR 识别失败: %s", e)
        raise HTTPException(status_code=503, detail=str(e))
    except Exception as e:
        logger.exception("OCR 识别失败")
        raise HTTPException(
            status_code=500,
            detail="AI识别服务暂时不可用，请稍后重试",
        )

    # 检查是否为孕期相关报告
    if not ocr_result.get("is_pregnancy_report", True):
        reject_reason = ocr_result.get("reject_reason", "")
        raise HTTPException(
            status_code=422,
            detail=f"请上传孕期检查报告。{reject_reason}",
        )

    # 根据报告日期 + 末次经期推算检查时的孕周
    report_date_str = ocr_result.get("report_date")
    pregnancy_week_at_report = user.pregnancy_week  # 默认用用户当前孕周
    if report_date_str and user.last_period_date:
        try:
            report_dt = date.fromisoformat(report_date_str)
            lmp = date.fromisoformat(user.last_period_date) if isinstance(user.last_period_date, str) else user.last_period_date
            delta_days = (report_dt - lmp).days
            if 0 < delta_days < 300:
                pregnancy_week_at_report = delta_days // 7
        except (ValueError, TypeError):
            pass

    # Create report with OCR data
    data_items = [DataItem(**item) for item in ocr_result.get("data_items", [])]
    report = Report(
        user_id=user.id,
        report_type=ocr_result.get("report_type", ""),
        pregnancy_week=pregnancy_week_at_report,
        report_date=report_date_str,
        image_urls=[url],
        data_items=data_items,
        patient_name=ocr_result.get("patient_name"),
        status="pending",
    )
    await report.insert()

    return {
        "report_id": str(report.id),
        "image_url": url,
        "ocr_result": {
            "report_type": report.report_type,
            "data_items": [item.model_dump() for item in data_items],
            "patient_name": mask_name(report.patient_name),
        },
    }


@router.post("/{report_id}/add-image")
async def add_image_to_report(
    report_id: str,
    file: UploadFile = File(...),
    user: User = Depends(get_current_user),
):
    """向已有报告追加图片并进行 OCR 识别，合并数据"""
    oid = parse_object_id(report_id)
    report = await Report.get(oid)
    if not report or report.user_id != user.id:
        raise HTTPException(status_code=404, detail="报告不存在")

    if file.content_type not in ALLOWED_CONTENT_TYPES:
        raise HTTPException(
            status_code=400,
            detail=f"不支持的文件类型: {file.content_type}，请上传 JPG/PNG/WebP 图片",
        )

    content = await file.read()
    if len(content) > MAX_FILE_SIZE:
        raise HTTPException(status_code=400, detail="文件大小不能超过 10MB")
    await file.seek(0)

    url = await save_upload_file(file, sub_dir="reports")

    # OCR recognition
    try:
        ocr_result = await ocr_recognize(url)
    except FileNotFoundError:
        raise HTTPException(status_code=500, detail="文件保存异常，请重试")
    except RuntimeError as e:
        logger.warning("OCR 识别失败: %s", e)
        raise HTTPException(status_code=503, detail=str(e))
    except Exception:
        logger.exception("OCR 识别失败")
        raise HTTPException(status_code=500, detail="AI识别服务暂时不可用，请稍后重试")

    # 检查是否为孕期相关报告
    if not ocr_result.get("is_pregnancy_report", True):
        reject_reason = ocr_result.get("reject_reason", "")
        raise HTTPException(
            status_code=422,
            detail=f"请上传孕期检查报告。{reject_reason}",
        )

    # Merge new data into existing report
    new_items = [DataItem(**item) for item in ocr_result.get("data_items", [])]
    report.image_urls.append(url)
    report.data_items.extend(new_items)

    # Update report_type if current report has multiple types
    new_type = ocr_result.get("report_type", "")
    if new_type and new_type != report.report_type:
        report.report_type = f"{report.report_type}+{new_type}" if report.report_type else new_type

    report.updated_at = datetime.now(timezone.utc)
    await report.save()

    return {
        "report_id": str(report.id),
        "image_url": url,
        "image_count": len(report.image_urls),
        "ocr_result": {
            "report_type": new_type,
            "data_items": [item.model_dump() for item in new_items],
            "patient_name": mask_name(ocr_result.get("patient_name")),
        },
        "total_items": len(report.data_items),
    }


@router.put("/{report_id}/verify", response_model=ReportResponse)
async def verify_report(
    report_id: str,
    req: ReportVerifyRequest,
    user: User = Depends(get_current_user),
):
    """数据校对确认"""
    oid = parse_object_id(report_id)
    report = await Report.get(oid)
    if not report or report.user_id != user.id:
        raise HTTPException(status_code=404, detail="报告不存在")

    report.data_items = [DataItem(**item.model_dump()) for item in req.data_items]
    report.symptoms = req.symptoms
    report.notes = req.notes
    if req.report_type:
        report.report_type = req.report_type
    report.status = "verified"

    # Calculate overall status
    statuses = [item.status for item in report.data_items]
    if any(s in ("high", "low") for s in statuses):
        abnormal_count = sum(1 for s in statuses if s != "normal")
        report.overall_status = "warning" if abnormal_count >= 3 else "attention"
    else:
        report.overall_status = "normal"

    report.updated_at = datetime.now(timezone.utc)
    await report.save()

    return _report_to_response(report)


@router.get("/list", response_model=list[ReportListItem])
async def list_reports(
    limit: int = 50,
    user: User = Depends(get_current_user),
):
    """获取报告列表"""
    reports = await Report.find(
        Report.user_id == user.id
    ).sort("-created_at").limit(limit).to_list()

    return [
        ReportListItem(
            id=str(r.id),
            report_type=r.report_type,
            pregnancy_week=r.pregnancy_week,
            overall_status=r.overall_status,
            summary=_generate_summary(r),
            item_count=len(r.data_items),
            created_at=r.created_at,
        )
        for r in reports
    ]


@router.get("/{report_id}", response_model=ReportResponse)
async def get_report(report_id: str, user: User = Depends(get_current_user)):
    """获取报告详情"""
    oid = parse_object_id(report_id)
    report = await Report.get(oid)
    if not report or report.user_id != user.id:
        raise HTTPException(status_code=404, detail="报告不存在")
    return _report_to_response(report)


def _generate_summary(report: Report) -> str:
    """从报告指标中生成有意义的摘要文本"""
    if not report.data_items:
        return "暂无指标数据"

    # 优先展示异常指标
    abnormal = [
        item for item in report.data_items if item.status in ("high", "low")
    ]
    # 再补充正常的关键指标
    normal = [item for item in report.data_items if item.status == "normal"]

    parts = []
    # 先添加异常指标（最多2个）
    for item in abnormal[:2]:
        flag = "↑偏高" if item.status == "high" else "↓偏低"
        val = f"{item.name}{item.value}{item.unit}（{flag}）" if item.unit else f"{item.name}{item.value}（{flag}）"
        parts.append(val)
    # 再补充正常指标，总共最多3个
    for item in normal[:3 - len(parts)]:
        val = f"{item.name}{item.value}{item.unit}" if item.unit else f"{item.name}{item.value}"
        parts.append(val)

    if not parts:
        return f"共{len(report.data_items)}项指标"

    return "，".join(parts)


def _report_to_response(report: Report) -> ReportResponse:
    return ReportResponse(
        id=str(report.id),
        report_type=report.report_type,
        pregnancy_week=report.pregnancy_week,
        image_urls=report.image_urls,
        data_items=[DataItemSchema(**item.model_dump()) for item in report.data_items],
        symptoms=report.symptoms,
        notes=report.notes,
        status=report.status,
        overall_status=report.overall_status,
        created_at=report.created_at,
    )