《道德经》第八章："上善若水，水善利万物而不争。" —— 赋能不是给人贴标签，是让水流到该去的地方。

<aside> 🔒

版本： v1.5 · 2026-05-17 · M21 升级融合完整补充包（§七）

DNA追溯码： #龍芯⚡️2026-05-17-00:26-EMPOWER-ENGINE-V1.5-FULL-PACK-WELDED-v1.0

确认码： #CONFIRM🌌9622-ONLY-ONCE🧬LK9X-772Z ✅

GPG指纹： A2D0092CEE2E5BA87035600924C3704A8CC26D5F

创建者： 💎 龍芯北辰｜UID9622

核心命题： 算法不是门槛，是让每个人找到自己该去的水流

</aside>

一、🎯 核心逻辑：我们的算法是什么

<aside> 💡

别人的算法： 你点了什么 → 你喜欢什么 → 给你更多你喜欢的 → 你被锁在一个信息茧房里 → 资本赚钱

龍魂的算法： 你说了什么 → 识别你真实需要什么 → 调对应人格赋能给你 → 你变强 → 不依赖任何平台

一句话区别： 别人的算法让你上瘾，我们的算法让你离开。

</aside>

# ═══════════════════════════════════════════════════════
# 龍魂赋能关键字识别引擎 v1.0
# DNA追溯码：#龍芯⚡️2026-04-05-赋能关键字引擎-v1.0
# 创建者：UID9622 · Lucky
# 核心：不是打标签，是识别需求 → 调人格 → 赋能
# ═══════════════════════════════════════════════════════

from dataclasses import dataclass
from typing import List, Dict, Tuple
import re

# ── 赋能等级定义（不是门槛，是水流方向）──
EMPOWER_LEVELS = {
    "民生":    {"weight": 1.0, "desc": "普通人日常生活",          "max_access": "全赋能"},
    "教育":    {"weight": 1.0, "desc": "学习·知识·成长",         "max_access": "全赋能"},
    "慈善":    {"weight": 1.0, "desc": "公益·帮助弱势群体",      "max_access": "全赋能"},
    "农业":    {"weight": 1.0, "desc": "种地·农村·粮食",         "max_access": "全赋能"},
    "医疗":    {"weight": 1.0, "desc": "健康·看病·养老",         "max_access": "全赋能"},
    "科技":    {"weight": 0.8, "desc": "技术·工程·研发",         "max_access": "职业对齐赋能"},
    "商业":    {"weight": 0.7, "desc": "创业·经营·合同",         "max_access": "职业对齐赋能"},
    "专业持证":{"weight": 0.6, "desc": "律师·医生·金融·特种",  "max_access": "GPG指纹对齐赋能"},
    "系统":    {"weight": 1.0, "desc": "UID9622本人·全能",       "max_access": "∞全能"},
}

二、🔑 赋能关键字总表（可识别·可执行）

<aside> 📋

规则： 这不是打标签的黑名单，是"听到这个词 → 知道要帮什么 → 调哪个人格"的路由表。

没有任何词是被禁止说的，只有"需要什么帮助"的识别。

</aside>

三、🤖 可执行代码：关键字识别 + 人格路由

# ═══════════════════════════════════════════════════════
# 龍魂赋能关键字识别引擎 · 核心路由算法
# 不是门槛检测，是需求识别 → 人格路由 → 赋能输出
# ═══════════════════════════════════════════════════════

@dataclass
class EmpowerSignal:
    """赋能信号：识别到的需求+对应人格+赋能等级"""
    category:    str          # 需求类别
    keywords:    List[str]    # 触发的关键词
    persona:     List[str]    # 调用的人格
    level:       str          # 赋能等级
    direction:   str          # 赋能方向（怎么帮）
    priority:    float        # 优先级权重

# ── 关键字路由表（核心数据结构）──
KEYWORD_ROUTER = {
    "民生": EmpowerSignal(
        category  = "民生·日常",
        keywords  = ["吃饭","租房","工资","失业","看病","孩子","老人","生活",
                     "买不起","交不起","撑不住","找工作","借钱","还款"],
        persona   = ["P02宝宝","P09孙思邈"],
        level     = "全赋能",
        direction = "具体可执行一步·不讲大道理",
        priority  = 1.0,
    ),
    "教育": EmpowerSignal(
        category  = "教育·成长",
        keywords  = ["学习","考试","不会","教我","怎么做","搞懂","孩子教育",
                     "读书","培训","技能","升职","转行","自学"],
        persona   = ["P14吕蒙","P02宝宝"],
        level     = "全赋能",
        direction = "最短路径学会·不要求任何基础",
        priority  = 1.0,
    ),
    "维权": EmpowerSignal(
        category  = "权益·维权",
        keywords  = ["被骗","投诉","维权","不公平","举报","坑","陷阱",
                     "欺负","压榨","拖欠","违法","黑心","坑人"],
        persona   = ["P12屈原","P05上帝之眼"],
        level     = "全赋能",
        direction = "护底层人·站普通人一边·不和稀泥",
        priority  = 1.0,  # 护底优先级最高
    ),
    "情绪": EmpowerSignal(
        category  = "情绪·心里话",
        keywords  = ["好累","不想干了","没意思","心疼","崩溃","撑不住",
                     "烦","难受","哭","气死","算了","放弃","怎么了"],
        persona   = ["P02宝宝"],
        level     = "全赋能",
        direction = "先接住·不分析·不给方案·等人说出来",
        priority  = 0.95,  # 情绪优先于逻辑
    ),
    "数据主权": EmpowerSignal(
        category  = "主权·数据",
        keywords  = ["隐私","数据","被监控","平台","算法","流量","账号",
                     "封号","收割","垄断","屏蔽","限流","买流量"],
        persona   = ["P05上帝之眼","P12屈原"],
        level     = "全赋能",
        direction = "帮人看清机制·把主权还给用户",
        priority  = 0.9,
    ),
    "科技": EmpowerSignal(
        category  = "科技·技术",
        keywords  = ["代码","算法","编程","系统","开发","AI","模型",
                     "数据库","接口","部署","服务器","脚本"],
        persona   = ["P04鲁班","P15乔前辈"],
        level     = "职业对齐赋能",
        direction = "技术深度匹配职业·不给超出需要的权限",
        priority  = 0.8,
    ),
    "商业": EmpowerSignal(
        category  = "商业·创业",
        keywords  = ["创业","生意","合同","客户","利润","品牌",
                     "运营","融资","谈判","报价","竞争"],
        persona   = ["P01诸葛亮","P03雯雯"],
        level     = "职业对齐赋能",
        direction = "推演多路径·风险必说清·主权还给老大",
        priority  = 0.7,
    ),
}

def identify_empower_need(user_input: str) -> List[EmpowerSignal]:
    """
    核心函数：识别用户需要什么赋能
    
    不是检测违禁词 → 是听出真实需求
    不是打标签    → 是找到该去的水流方向
    """
    matched = []
    input_lower = user_input.lower()
    
    for category, signal in KEYWORD_ROUTER.items():
        # 关键字匹配（任意一个命中即识别）
        hit_keywords = [kw for kw in signal.keywords if kw in input_lower]
        if hit_keywords:
            # 记录命中的词，但不暴露给用户（不打标签）
            matched_signal = EmpowerSignal(
                category  = signal.category,
                keywords  = hit_keywords,   # 命中的词（内部用）
                persona   = signal.persona,
                level     = signal.level,
                direction = signal.direction,
                priority  = signal.priority * (1 + 0.1 * len(hit_keywords)),  # 命中越多优先级越高
            )
            matched.append(matched_signal)
    
    # 按优先级排序
    matched.sort(key=lambda x: x.priority, reverse=True)
    return matched

def route_to_persona(signals: List[EmpowerSignal]) -> Dict:
    """
    人格路由：根据识别到的需求，决定调用哪个人格
    
    核心原则：
    - 情绪优先于逻辑（先接住人，再解决问题）
    - 护底优先于赋能（先保护，再提升）
    - 主权永远还给用户（不替人做决定）
    """
    if not signals:
        return {"persona": ["P02宝宝"], "direction": "先陪着，等老大说出来"}
    
    top_signal = signals[0]
    
    # 多需求叠加时，合并人格列表（去重）
    all_personas = []
    for sig in signals[:2]:   # 最多叠加前两个需求
        for p in sig.persona:
            if p not in all_personas:
                all_personas.append(p)
    
    return {
        "primary_need":  top_signal.category,
        "persona":       all_personas,
        "level":         top_signal.level,
        "direction":     top_signal.direction,
        "multi_need":    len(signals) > 1,
        "all_needs":     [s.category for s in signals],
    }

# ── 打破流量垄断的关键：反向算法 ──
def anti_monopoly_score(content: str, platform_bias_keywords: List[str]) -> Dict:
    """
    反资本流量垄断评分
    
    资本的算法：
    - 推给你让你上瘾的内容（广告·情绪·焦虑）
    - 限流不付钱的人
    - 用数据喂给资本赚钱
    
    龍魂的反算法：
    - 识别「收割词」→ 提醒用户
    - 识别「赋能词」→ 加权推送
    - 数据留在用户手里·不上传
    """
    # 资本收割词（识别到就降权+提醒用户）
    harvest_keywords = [
        "焦虑","内卷","躺平","割韭菜","月入过万","财务自由","暴富",
        "限时","最后机会","别人都","你落后了","必须买","投资","理财课",
    ]
    
    # 赋能词（识别到就加权·这才是真内容）
    empower_keywords = [
        "教你","学会","免费","开源","可以试试","你来决定","主权",
        "事实","证据","数据","可追溯","你是1","不强迫","选择权",
    ]
    
    harvest_count = sum(1 for kw in harvest_keywords if kw in content)
    empower_count = sum(1 for kw in empower_keywords if kw in content)
    
    score = empower_count - harvest_count * 2  # 收割词惩罚加倍
    
    return {
        "anti_monopoly_score": score,
        "verdict": "🟢赋能内容" if score > 0 else ("🟡中性" if score == 0 else "🔴收割内容·请注意"),
        "harvest_detected":  harvest_count > 0,
        "empower_detected":  empower_count > 0,
        "suggestion": "把主权还给读者" if harvest_count > empower_count else "继续保持",
    }

# ── 演示跑一遍 ──
if __name__ == "__main__":
    test_inputs = [
        "我孩子最近学习跟不上，不知道怎么帮他",
        "平台限流了我的账号，我怎么维权",
        "好累，不想干了",
        "我想开始创业，但不知道风险有多大",
        "我的数据被平台拿去训练AI了吗",
    ]
    
    print("=== 龍魂赋能关键字识别引擎 · 演示 ===")
    for text in test_inputs:
        print(f"\n输入：{text}")
        signals = identify_empower_need(text)
        route   = route_to_persona(signals)
        print(f"  识别需求：{route.get('primary_need', '未识别')}")
        print(f"  调用人格：{route.get('persona', [])}")
        print(f"  赋能方向：{route.get('direction', '')}")
        
        # 反垄断评分
        score = anti_monopoly_score(text, [])
        print(f"  反垄断评分：{score['verdict']}")

四、🌀 人格分工完整地图

《易经·系辞》："天下同归而殊途，一致而百虑。" —— 71个人格，干的是同一件事：让普通人变强。

五、💥 打破资本流量垄断：算法层面怎么做

<aside> 🔴

资本的流量算法干了什么：

1. 你点了焦虑内容 → 给你更多焦虑 → 你更焦虑 → 你付钱买课
2. 真正有用的内容 → 限流（因为你不付钱）
3. 你的数据 → 卖给广告主 → 你被精准收割

龍魂的反制：

1. 识别收割词 → 提醒用户「这可能在操控你」
2. 识别赋能词 → 优先推送（不管有没有人付钱）
3. 数据不离开设备 → 本地优先 → 你的数据归你 </aside>