"""垃圾信息分类标签模块

在朴素贝叶斯二分类（spam/ham）基础上，对判定为 spam 的文本进行细分类标签。
分类优先级：诈骗 > 骚扰 > 广告（按危害程度排序）
"""

CATEGORY_KEYWORDS = {
    "fraud": [
        "中奖", "幸运粉丝", "幸运用户", "银行卡异常", "社保异常", "账号冻结",
        "解封", "立即验证", "验证码", "欠费停机", "退款待确认", "违章信息",
        "紧急通知", "账户异常", "风险", "核验", "被冻结", "将被冻结",
    ],
    "harassment": [
        "兼职", "日结", "高薪", "刷单", "赚钱", "外快", "宝妈", "学生都能做",
        "添加微信", "扫码进群", "进群立刻", "想赚", "零花钱", "在家办公",
        "无需面试", "火热招募", "秒赚", "招募",
    ],
    "advertisement": [
        "领取", "优惠", "红包", "优惠券", "秒杀", "返现", "补贴", "会员",
        "特价", "低价", "点击链接", "扫码", "免费领取", "无门槛", "现金券",
        "盲盒", "百分百中奖", "隐藏优惠券", "内部价", "货到付款", "限时",
        "最后", "名额", "先到先得",
    ],
}

CATEGORY_LABELS = {
    "fraud": "疑似诈骗",
    "harassment": "疑似骚扰",
    "advertisement": "疑似广告",
    "spam": "疑似垃圾",
    "ham": "",
}

CATEGORY_PRIORITY = ["fraud", "harassment", "advertisement"]


def categorize_spam(text: str) -> tuple[str, str]:
    """根据关键词匹配判定垃圾信息的具体分类标签

    Args:
        text: 待分类的文本内容

    Returns:
        tuple[str, str]: (category_code, category_label)
            - category_code: fraud | harassment | advertisement | spam
            - category_label: 疑似诈骗 | 疑似骚扰 | 疑似广告 | 疑似垃圾
    """
    text_lower = text.lower()

    for category in CATEGORY_PRIORITY:
        keywords = CATEGORY_KEYWORDS.get(category, [])
        for kw in keywords:
            if kw.lower() in text_lower:
                return category, CATEGORY_LABELS[category]

    return "spam", CATEGORY_LABELS["spam"]


def get_category_label(category: str) -> str:
    """获取分类标签的中文显示文本

    Args:
        category: 分类代码

    Returns:
        str: 中文标签文本
    """
    return CATEGORY_LABELS.get(category, "")