DDUp/src/vitals/vision/analyzer.py

"""食物图片识别分析器"""

import base64
from pathlib import Path
from typing import Optional

from ..core.calories import estimate_meal_calories


class FoodAnalyzer:
    """食物分析器基类"""

    def analyze(self, image_path: Path) -> dict:
        """
        分析食物图片

        返回:
        {
            "description": "米饭、红烧排骨、西兰花",
            "total_calories": 680,
            "total_protein": 25.0,
            "total_carbs": 85.0,
            "total_fat": 20.0,
            "items": [...]
        }
        """
        raise NotImplementedError


class ClaudeFoodAnalyzer(FoodAnalyzer):
    """使用 Claude Vision API 的食物分析器"""

    def __init__(self, api_key: Optional[str] = None):
        from ..core import database as db
        self.api_key = api_key or db.get_api_key("anthropic")

    def analyze(self, image_path: Path) -> dict:
        """使用 Claude Vision 分析食物图片"""
        if not self.api_key:
            raise ValueError("需要设置 ANTHROPIC_API_KEY 环境变量")

        from anthropic import Anthropic

        client = Anthropic(api_key=self.api_key)

        # 读取并编码图片
        image_data = self._encode_image(image_path)
        media_type = self._get_media_type(image_path)

        # 调用 Claude Vision API
        message = client.messages.create(
            model="claude-sonnet-4-20250514",
            max_tokens=1024,
            messages=[
                {
                    "role": "user",
                    "content": [
                        {
                            "type": "image",
                            "source": {
                                "type": "base64",
                                "media_type": media_type,
                                "data": image_data,
                            },
                        },
                        {
                            "type": "text",
                            "text": """请分析这张食物图片，识别出所有食物。

要求：
1. 列出所有可识别的食物，用中文名称
2. 估计每种食物的大致份量
3. 按照 "食物名称" 的格式返回，多个食物用加号连接

例如返回格式：米饭+红烧排骨+西兰花

只返回食物列表，不需要其他解释。"""
                        },
                    ],
                }
            ],
        )

        # 解析响应
        description = message.content[0].text.strip()

        # 使用卡路里计算模块估算营养成分
        result = estimate_meal_calories(description)
        result["description"] = description

        return result

    def _encode_image(self, image_path: Path) -> str:
        """将图片编码为 base64"""
        with open(image_path, "rb") as f:
            return base64.standard_b64encode(f.read()).decode("utf-8")

    def _get_media_type(self, image_path: Path) -> str:
        """获取图片的 MIME 类型"""
        suffix = image_path.suffix.lower()
        media_types = {
            ".jpg": "image/jpeg",
            ".jpeg": "image/jpeg",
            ".png": "image/png",
            ".gif": "image/gif",
            ".webp": "image/webp",
        }
        return media_types.get(suffix, "image/jpeg")


class LocalFoodAnalyzer(FoodAnalyzer):
    """本地食物分析器（简单实现，基于文件名）"""

    def analyze(self, image_path: Path) -> dict:
        """
        简单的本地分析器
        实际使用时可以替换为本地 AI 模型
        """
        # 目前仅返回空结果，提示用户手动输入
        return {
            "description": "",
            "total_calories": 0,
            "total_protein": 0,
            "total_carbs": 0,
            "total_fat": 0,
            "items": [],
            "note": "本地分析器暂不支持图片识别，请手动输入食物描述",
        }


def get_analyzer(
    provider: str = "qwen",
    api_key: Optional[str] = None,
    use_claude: bool = False,  # 保留向后兼容
) -> FoodAnalyzer:
    """
    获取食物分析器

    provider: "qwen" | "deepseek" | "claude" | "local"
    """
    # 向后兼容: 如果使用旧参数 use_claude=True
    if use_claude:
        provider = "claude"

    if provider == "qwen":
        from .providers.qwen import get_qwen_analyzer
        return get_qwen_analyzer(api_key)
    elif provider == "deepseek":
        from .providers.deepseek import get_deepseek_analyzer
        return get_deepseek_analyzer(api_key)
    elif provider == "claude":
        return ClaudeFoodAnalyzer(api_key)
    else:
        return LocalFoodAnalyzer()