"""阿里云 OCR 客户端封装（ocr-api 2021-07-07）。

只关注当前需要的能力——RecognizePassport（护照识别）。
其它证件接口（身份证、行驶证…）按需扩展即可，签名/客户端复用 _build_client。

配置项 (app.config.settings):
- aliyun_ocr_access_key_id
- aliyun_ocr_access_key_secret
- aliyun_ocr_endpoint  (默认 ocr-api.cn-hangzhou.aliyuncs.com)
"""
from __future__ import annotations

import json
from functools import lru_cache
from io import BytesIO
from typing import Any

from app.config import settings


class OcrConfigError(RuntimeError):
    """AK/SK 未配置等可恢复错误，路由层应返回 503。"""


class OcrCallError(RuntimeError):
    """阿里云返回非 200 或网络异常等运行期错误。"""


@lru_cache(maxsize=1)
def _build_client():  # 延迟导入 SDK，缺包时给出明确提示
    try:
        from alibabacloud_ocr_api20210707.client import Client
        from alibabacloud_tea_openapi.models import Config
    except ImportError as exc:  # noqa: BLE001
        raise OcrConfigError(
            "未安装阿里云 OCR SDK，请执行: pip install alibabacloud_ocr_api20210707"
        ) from exc

    ak = settings.aliyun_ocr_access_key_id.strip()
    sk = settings.aliyun_ocr_access_key_secret.strip()
    endpoint = settings.aliyun_ocr_endpoint.strip() or "ocr-api.cn-hangzhou.aliyuncs.com"
    if not ak or not sk:
        raise OcrConfigError(
            "阿里云 OCR 未配置：请在 .env 设置 ALIYUN_OCR_ACCESS_KEY_ID / ALIYUN_OCR_ACCESS_KEY_SECRET"
        )
    config = Config(access_key_id=ak, access_key_secret=sk)
    config.endpoint = endpoint
    return Client(config)


def _parse_data(body: Any) -> dict[str, Any]:
    """RecognizePassport 返回 body.data 是一段 JSON 字符串，统一解析为 dict。"""
    raw = getattr(body, "data", None)
    if raw is None and isinstance(body, dict):
        raw = body.get("Data") or body.get("data")
    if raw is None:
        return {}
    if isinstance(raw, (dict, list)):
        return raw  # type: ignore[return-value]
    try:
        return json.loads(raw)
    except (TypeError, ValueError):
        return {"raw": str(raw)}


def recognize_passport(
    *,
    image_bytes: bytes | None = None,
    image_url: str | None = None,
) -> dict[str, Any]:
    """识别护照。image_bytes / image_url 至少传一个。

    返回结构（与控制台字段一致，常见字段示例）:
        {
            "name": "...",                # 姓名
            "sex": "M",
            "country": "CHN",
            "passport_no": "E12345678",
            "date_of_birth": "1990-01-01",
            "date_of_expiry": "2030-01-01",
            "place_of_birth": "...",
            ...其它字段透传...
            "_request_id": "...",         # 便于在阿里云后台溯源
        }
    """
    if not image_bytes and not image_url:
        raise ValueError("recognize_passport 需要 image_bytes 或 image_url")

    from alibabacloud_ocr_api20210707.models import RecognizePassportRequest
    from alibabacloud_tea_util.models import RuntimeOptions

    client = _build_client()
    req_kwargs: dict[str, Any] = {}
    if image_bytes is not None:
        req_kwargs["body"] = BytesIO(image_bytes)
    if image_url:
        req_kwargs["url"] = image_url
    request = RecognizePassportRequest(**req_kwargs)

    try:
        response = client.recognize_passport_with_options(request, RuntimeOptions())
    except OcrConfigError:
        raise
    except Exception as exc:  # noqa: BLE001
        raise OcrCallError(f"调用阿里云 OCR 失败: {exc}") from exc

    body = getattr(response, "body", response)
    data = _parse_data(body)
    request_id = getattr(body, "request_id", None) or (
        body.get("RequestId") if isinstance(body, dict) else None
    )
    if request_id and isinstance(data, dict):
        data.setdefault("_request_id", request_id)
    return data